Google Git
Sign in
apache/impala/refs/heads/master/./be/src/exec/hdfs-scan-node-base.cc
blob: 961faa501f001169529fe5f93be1c283973163c1 [file]
// Licensed to the Apache Software Foundation (ASF) under one
// or more contributor license agreements. See the NOTICE file
// distributed with this work for additional information
// regarding copyright ownership. The ASF licenses this file
// to you under the Apache License, Version 2.0 (the
// "License"); you may not use this file except in compliance
// with the License. You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing,
// software distributed under the License is distributed on an
// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
// KIND, either express or implied. See the License for the
// specific language governing permissions and limitations
// under the License.
#include "exec/hdfs-scan-node-base.h"
#include "exec/base-sequence-scanner.h"
#include "exec/hdfs-columnar-scanner.h"
#include "exec/hdfs-scan-node-mt.h"
#include "exec/hdfs-scan-node.h"
#include "exec/avro/hdfs-avro-scanner.h"
#include "exec/orc/hdfs-orc-scanner.h"
#include "exec/parquet/hdfs-parquet-scanner.h"
#include "exec/rcfile/hdfs-rcfile-scanner.h"
#include "exec/sequence/hdfs-sequence-scanner.h"
#include "exec/text/hdfs-text-scanner.h"
#include "exec/text/hdfs-plugin-text-scanner.h"
#include "exec/json/hdfs-json-scanner.h"
#include <avro/errors.h>
#include <avro/schema.h>
#include <boost/filesystem.hpp>
#include <gutil/strings/substitute.h>
#include "codegen/llvm-codegen.h"
#include "common/logging.h"
#include "common/object-pool.h"
#include "exprs/scalar-expr-evaluator.h"
#include "exprs/scalar-expr.h"
#include "runtime/descriptors.h"
#include "runtime/exec-env.h"
#include "runtime/fragment-state.h"
#include "runtime/hdfs-fs-cache.h"
#include "runtime/io/disk-io-mgr.h"
#include "runtime/io/request-context.h"
#include "runtime/query-state.h"
#include "runtime/runtime-filter.inline.h"
#include "runtime/runtime-state.h"
#include "util/compression-util.h"
#include "util/disk-info.h"
#include "util/flat_buffer.h"
#include "util/hdfs-util.h"
#include "util/impalad-metrics.h"
#include "util/metrics.h"
#include "util/periodic-counter-updater.h"
#include "util/pretty-printer.h"
#include "util/scope-exit-trigger.h"
#include "common/names.h"
#ifndef NDEBUG
DECLARE_bool(skip_file_runtime_filtering);
#endif
DEFINE_bool(always_use_data_cache, false, "(Advanced) Always uses the IO data cache "
"for all reads, regardless of whether the read is local or remote. By default, the "
"IO data cache is only used if the data is expected to be remote. Used by tests.");
namespace filesystem = boost::filesystem;
using namespace impala::io;
using namespace strings;
namespace impala {
PROFILE_DEFINE_TIMER(TotalRawHdfsReadTime, STABLE_LOW, "Aggregate wall clock time"
" across all Disk I/O threads in HDFS read operations.");
PROFILE_DEFINE_TIMER(TotalRawHdfsOpenFileTime, STABLE_LOW, "Aggregate wall clock time"
" spent across all Disk I/O threads in HDFS open operations.");
PROFILE_DEFINE_DERIVED_COUNTER(PerReadThreadRawHdfsThroughput, STABLE_LOW,
TUnit::BYTES_PER_SECOND, "The read throughput in bytes/sec for each HDFS read thread"
" while it is executing I/O operations on behalf of a scan.");
PROFILE_DEFINE_COUNTER(ScanRangesComplete, STABLE_LOW, TUnit::UNIT,
"Number of scan ranges that have been completed by a scan node.");
PROFILE_DEFINE_COUNTER(CollectionItemsRead, STABLE_LOW, TUnit::UNIT,
"Total number of nested collection items read by the scan. Only created for scans "
"(e.g. Parquet) that support nested types.");
PROFILE_DEFINE_COUNTER(NumDisksAccessed, STABLE_LOW, TUnit::UNIT, "Number of distinct "
"disks accessed by HDFS scan. Each local disk is counted as a disk and each type of"
" remote filesystem (e.g. HDFS remote reads, S3) is counted as a distinct disk.");
PROFILE_DEFINE_SAMPLING_COUNTER(AverageHdfsReadThreadConcurrency, STABLE_LOW, "The"
" average number of HDFS read threads executing read operations on behalf of this "
"scan. Higher values (i.e. close to the aggregate number of I/O threads across "
"all disks accessed) show that this scan is using a larger proportion of the I/O "
"capacity of the system. Lower values show that either this scan is not I/O bound"
" or that it is getting a small share of the I/O capacity of the system.");
PROFILE_DEFINE_SUMMARY_STATS_COUNTER(InitialRangeIdealReservation, DEBUG, TUnit::BYTES,
"Tracks stats about the ideal reservation for initial scan ranges. Use this to "
"determine if the scan got all of the reservation it wanted. Does not include "
"subsequent reservation increases done by scanner implementation (e.g. for Parquet "
"columns).");
PROFILE_DEFINE_SUMMARY_STATS_COUNTER(InitialRangeActualReservation, DEBUG,
TUnit::BYTES, "Tracks stats about the actual reservation for initial scan ranges. "
"Use this to determine if the scan got all of the reservation it wanted. Does not "
"include subsequent reservation increases done by scanner implementation "
"(e.g. for Parquet columns).");
PROFILE_DEFINE_COUNTER(BytesReadLocal, STABLE_LOW, TUnit::BYTES,
"The total number of bytes read locally");
PROFILE_DEFINE_COUNTER(BytesReadShortCircuit, STABLE_LOW, TUnit::BYTES,
"The total number of bytes read via short circuit read");
PROFILE_DEFINE_COUNTER(BytesReadDataNodeCache, STABLE_HIGH, TUnit::BYTES,
"The total number of bytes read from data node cache");
PROFILE_DEFINE_COUNTER(BytesReadEncrypted, STABLE_LOW, TUnit::BYTES,
"The total number of bytes read from encrypted data");
PROFILE_DEFINE_COUNTER(BytesReadErasureCoded, STABLE_LOW, TUnit::BYTES,
"The total number of bytes read from erasure-coded data");
PROFILE_DEFINE_COUNTER(RemoteScanRanges, STABLE_HIGH, TUnit::UNIT,
"The total number of remote scan ranges");
PROFILE_DEFINE_COUNTER(BytesReadRemoteUnexpected, STABLE_LOW, TUnit::BYTES,
"The total number of bytes read remotely that were expected to be local");
PROFILE_DEFINE_COUNTER(CachedFileHandlesHitCount, STABLE_LOW, TUnit::UNIT,
"Total number of file handle opens where the file handle was present in the cache");
PROFILE_DEFINE_COUNTER(CachedFileHandlesMissCount, STABLE_LOW, TUnit::UNIT,
"Total number of file handle opens where the file handle was not in the cache");
PROFILE_DEFINE_HIGH_WATER_MARK_COUNTER(MaxCompressedTextFileLength, STABLE_LOW,
TUnit::BYTES, "The size of the largest compressed text file to be scanned. "
"This is used to estimate scanner thread memory usage.");
PROFILE_DEFINE_TIMER(ScannerIoWaitTime, STABLE_LOW, "Total amount of time scanner "
"threads spent waiting for I/O. This value can be compared to the value of "
"ScannerThreadsTotalWallClockTime of MT_DOP = 0 scan nodes or otherwise compared "
"to the total time reported for MT_DOP > 0 scan nodes. High values show that "
"scanner threads are spending significant time waiting for I/O instead of "
"processing data. Note that this includes the time when the thread is runnable "
"but not scheduled.");
PROFILE_DEFINE_SUMMARY_STATS_COUNTER(ParquetUncompressedBytesReadPerColumn, STABLE_LOW,
TUnit::BYTES, "Stats about the number of uncompressed bytes read per column. "
"Each sample in the counter is the size of a single column that is scanned by the "
"scan node.");
PROFILE_DEFINE_SUMMARY_STATS_COUNTER(ParquetCompressedBytesReadPerColumn, STABLE_LOW,
TUnit::BYTES, "Stats about the number of compressed bytes read per column. "
"Each sample in the counter is the size of a single column that is scanned by the "
"scan node.");
PROFILE_DEFINE_COUNTER(DataCacheHitCount, STABLE_HIGH, TUnit::UNIT,
"Total count of data cache hit");
PROFILE_DEFINE_COUNTER(DataCachePartialHitCount, STABLE_HIGH, TUnit::UNIT,
"Total count of data cache partially hit");
PROFILE_DEFINE_COUNTER(DataCacheMissCount, STABLE_HIGH, TUnit::UNIT,
"Total count of data cache miss");
PROFILE_DEFINE_COUNTER(DataCacheHitBytes, STABLE_HIGH, TUnit::BYTES,
"Total bytes of data cache hit");
PROFILE_DEFINE_COUNTER(DataCacheMissBytes, STABLE_HIGH, TUnit::BYTES,
"Total bytes of data cache miss");
const string HdfsScanNodeBase::HDFS_SPLIT_STATS_DESC =
"Hdfs split stats (<volume id>:<# splits>/<split lengths>)";
// Determines how many unexpected remote bytes trigger an error in the runtime state
const int UNEXPECTED_REMOTE_BYTES_WARN_THRESHOLD = 64 * 1024 * 1024;
Status HdfsScanPlanNode::Init(const TPlanNode& tnode, FragmentState* state) {
RETURN_IF_ERROR(ScanPlanNode::Init(tnode, state));
tuple_id_ = tnode.hdfs_scan_node.tuple_id;
tuple_desc_ = state->desc_tbl().GetTupleDescriptor(tuple_id_);
DCHECK(tuple_desc_->table_desc() != NULL);
hdfs_table_ = static_cast<const HdfsTableDescriptor*>(tuple_desc_->table_desc());
// Parse Avro table schema if applicable
const string& avro_schema_str = hdfs_table_->avro_schema();
if (!avro_schema_str.empty()) {
avro_schema_t avro_schema;
int error = avro_schema_from_json_length(
avro_schema_str.c_str(), avro_schema_str.size(), &avro_schema);
if (error != 0) {
return Status(Substitute("Failed to parse table schema: $0", avro_strerror()));
}
RETURN_IF_ERROR(AvroSchemaElement::ConvertSchema(avro_schema, avro_schema_.get()));
}
// Gather materialized partition-key slots and non-partition slots.
const vector<SlotDescriptor*>& slots = tuple_desc_->slots();
for (size_t i = 0; i < slots.size(); ++i) {
if (UNLIKELY(slots[i]->IsVirtual())) {
virtual_column_slots_.push_back(slots[i]);
} else if (hdfs_table_->IsClusteringCol(slots[i])) {
partition_key_slots_.push_back(slots[i]);
} else {
materialized_slots_.push_back(slots[i]);
}
}
// Order the materialized slots such that for schemaless file formats (e.g. text) the
// order corresponds to the physical order in files. For formats where the file schema
// is independent of the table schema (e.g. Avro, Parquet), this step is not necessary.
sort(materialized_slots_.begin(), materialized_slots_.end(),
SlotDescriptor::ColPathLessThan);
// Populate mapping from slot path to index into materialized_slots_.
for (int i = 0; i < materialized_slots_.size(); ++i) {
path_to_materialized_slot_idx_[materialized_slots_[i]->col_path()] = i;
}
// Initialize is_materialized_col_
is_materialized_col_.resize(hdfs_table_->num_cols());
for (int i = 0; i < hdfs_table_->num_cols(); ++i) {
is_materialized_col_[i] =
GetMaterializedSlotIdx(vector<int>(1, i)) != HdfsScanNodeBase::SKIP_COLUMN;
}
// Add collection item conjuncts
for (const auto& entry : tnode.hdfs_scan_node.collection_conjuncts) {
TupleDescriptor* tuple_desc = state->desc_tbl().GetTupleDescriptor(entry.first);
RowDescriptor* collection_row_desc =
state->obj_pool()->Add(new RowDescriptor(tuple_desc, /* is_nullable */ false));
DCHECK(conjuncts_map_.find(entry.first) == conjuncts_map_.end());
RETURN_IF_ERROR(ScalarExpr::Create(
entry.second, *collection_row_desc, state, &conjuncts_map_[entry.first]));
}
const TTupleId& tuple_id = tnode.hdfs_scan_node.tuple_id;
DCHECK(conjuncts_map_[tuple_id].empty());
conjuncts_map_[tuple_id] = conjuncts_;
// Add stats conjuncts
if (tnode.hdfs_scan_node.__isset.stats_tuple_id) {
TupleDescriptor* stats_tuple_desc =
state->desc_tbl().GetTupleDescriptor(tnode.hdfs_scan_node.stats_tuple_id);
DCHECK(stats_tuple_desc != nullptr);
RowDescriptor* stats_row_desc = state->obj_pool()->Add(
new RowDescriptor(stats_tuple_desc, /* is_nullable */ false));
RETURN_IF_ERROR(ScalarExpr::Create(tnode.hdfs_scan_node.stats_conjuncts,
*stats_row_desc, state, &stats_conjuncts_));
}
// Transfer overlap predicate descs.
overlap_predicate_descs_ = tnode.hdfs_scan_node.overlap_predicate_descs;
RETURN_IF_ERROR(ProcessScanRangesAndInitSharedState(state));
state->CheckAndAddCodegenDisabledMessage(codegen_status_msgs_);
return Status::OK();
}
Status HdfsScanPlanNode::ProcessScanRangesAndInitSharedState(FragmentState* state) {
// Initialize the template tuple pool.
using namespace org::apache::impala::fb;
shared_state_.template_pool_.reset(new MemPool(state->query_mem_tracker()));
auto& template_tuple_map_ = shared_state_.partition_template_tuple_map_;
ObjectPool* obj_pool = shared_state_.obj_pool();
auto& file_descs = shared_state_.file_descs_;
HdfsFsCache::HdfsFsMap fs_cache;
int num_ranges_missing_volume_id = 0;
int64_t total_splits = 0;
const vector<const PlanFragmentInstanceCtxPB*>& instance_ctx_pbs =
state->instance_ctx_pbs();
auto instance_ctxs = state->instance_ctxs();
DCHECK_EQ(instance_ctxs.size(), instance_ctx_pbs.size());
for (int i = 0; i < instance_ctxs.size(); ++i) {
auto ctx = instance_ctx_pbs[i];
auto instance_ctx = instance_ctxs[i];
auto ranges = ctx->per_node_scan_ranges().find(tnode_->node_id);
if (ranges == ctx->per_node_scan_ranges().end()) continue;
for (const ScanRangeParamsPB& params : ranges->second.scan_ranges()) {
DCHECK(params.scan_range().has_hdfs_file_split());
const HdfsFileSplitPB& split = params.scan_range().hdfs_file_split();
const org::apache::impala::fb::FbSplitFileMetadata* file_metadata = nullptr;
if (params.scan_range().has_file_metadata()) {
file_metadata =
flatbuffers::GetRoot<org::apache::impala::fb::FbSplitFileMetadata>(
params.scan_range().file_metadata().c_str());
}
HdfsPartitionDescriptor* partition_desc =
hdfs_table_->GetPartition(split.partition_id());
if (template_tuple_map_.find(split.partition_id()) == template_tuple_map_.end()) {
template_tuple_map_[split.partition_id()] =
InitTemplateTuple(partition_desc->partition_key_value_evals(),
shared_state_.template_pool_.get());
}
// Convert the ScanRangeParamsPB into per-file DiskIO::ScanRange objects and
// populate partition_ids_, file_descs_, and per_type_files_.
if (partition_desc == nullptr) {
// TODO: this should be a DCHECK but we sometimes hit it. It's likely IMPALA-1702.
LOG(ERROR) << "Bad table descriptor! table_id=" << hdfs_table_->id()
<< " partition_id=" << split.partition_id() << "\n"
<< state->fragment()
<< state->fragment_ctx().DebugString();
return Status("Query encountered invalid metadata, likely due to IMPALA-1702."
" Try rerunning the query.");
}
filesystem::path file_path;
if (hdfs_table_->IsIcebergTable() && split.relative_path().empty()) {
file_path.append(split.absolute_path(), filesystem::path::codecvt());
} else {
file_path.append(partition_desc->location(), filesystem::path::codecvt())
.append(split.relative_path(), filesystem::path::codecvt());
}
const string& native_file_path = file_path.native();
auto file_desc_map_key = make_pair(partition_desc->id(), native_file_path);
HdfsFileDesc* file_desc = nullptr;
auto file_desc_it = file_descs.find(file_desc_map_key);
if (file_desc_it == file_descs.end()) {
// Add new file_desc to file_descs_ and per_type_files_
file_desc = obj_pool->Add(new HdfsFileDesc(native_file_path));
file_descs[file_desc_map_key] = file_desc;
file_desc->file_length = split.file_length();
file_desc->mtime = split.mtime();
file_desc->file_compression = CompressionTypePBToThrift(split.file_compression());
file_desc->is_encrypted = split.is_encrypted();
file_desc->is_erasure_coded = split.is_erasure_coded();
file_desc->file_metadata = file_metadata;
file_desc->fragment_instance_id = instance_ctx->fragment_instance_id;
if (file_metadata) {
DCHECK(file_metadata->iceberg_metadata() != nullptr);
switch (file_metadata->iceberg_metadata()->file_format()) {
case FbIcebergDataFileFormat::FbIcebergDataFileFormat_PARQUET:
file_desc->file_format = THdfsFileFormat::PARQUET;
break;
case FbIcebergDataFileFormat::FbIcebergDataFileFormat_ORC:
file_desc->file_format = THdfsFileFormat::ORC;
break;
case FbIcebergDataFileFormat::FbIcebergDataFileFormat_AVRO:
file_desc->file_format = THdfsFileFormat::AVRO;
break;
case FbIcebergDataFileFormat::FbIcebergDataFileFormat_PUFFIN:
file_desc->file_format = THdfsFileFormat::PUFFIN;
break;
default:
return Status(Substitute(
"Unknown Iceberg file format type: $0",
file_metadata->iceberg_metadata()->file_format()));
}
} else {
file_desc->file_format = partition_desc->file_format();
}
RETURN_IF_ERROR(HdfsFsCache::instance()->GetConnection(
native_file_path, &file_desc->fs, &fs_cache));
shared_state_.per_type_files_[partition_desc->file_format()].push_back(file_desc);
} else {
// File already processed
file_desc = file_desc_it->second;
}
bool expected_local = params.has_is_remote() && !params.is_remote();
if (expected_local && params.volume_id() == -1) {
// IMPALA-11541 TODO: Ozone returns a list of null volume IDs. So we know the
// number of volumes, but ID will be -1. Skip this metric for Ozone paths because
// it doesn't convey useful feedback.
if (!IsOzonePath(partition_desc->location().c_str())) {
++num_ranges_missing_volume_id;
}
}
int cache_options = BufferOpts::NO_CACHING;
if (params.has_try_hdfs_cache() && params.try_hdfs_cache()) {
cache_options |= BufferOpts::USE_HDFS_CACHE;
}
if ((!expected_local || FLAGS_always_use_data_cache)
&& !state->query_options().disable_data_cache) {
cache_options |= BufferOpts::USE_DATA_CACHE;
}
ScanRangeMetadata* metadata =
obj_pool->Add(new ScanRangeMetadata(split.partition_id(), nullptr));
file_desc->splits.push_back(ScanRange::AllocateScanRange(obj_pool,
file_desc->GetFileInfo(), split.length(), split.offset(), {}, metadata,
params.volume_id(), expected_local, BufferOpts(cache_options)));
total_splits++;
}
// Update server wide metrics for number of scan ranges and ranges that have
// incomplete metadata.
ImpaladMetrics::NUM_RANGES_PROCESSED->Increment(ranges->second.scan_ranges().size());
ImpaladMetrics::NUM_RANGES_MISSING_VOLUME_ID->Increment(num_ranges_missing_volume_id);
}
// Set up the rest of the shared state.
shared_state_.remaining_scan_range_submissions_.Store(instance_ctx_pbs.size());
shared_state_.progress().Init(
Substitute("Splits complete (node=$0)", tnode_->node_id), total_splits);
shared_state_.use_mt_scan_node_ = tnode_->hdfs_scan_node.use_mt_scan_node;
shared_state_.scan_range_queue_.Reserve(total_splits);
// Distribute the work evenly for issuing initial scan ranges.
DCHECK(shared_state_.use_mt_scan_node_ || instance_ctx_pbs.size() == 1)
<< "Non MT scan node should only have a single instance.";
if (tnode_->hdfs_scan_node.deterministic_scanrange_assignment) {
// If using deterministic scan range assignment, there is no need to rebalance
// the scan ranges. The scan ranges stay with their original fragment instance.
for (auto& fd : file_descs) {
const TUniqueId& instance_id = fd.second->fragment_instance_id;
shared_state_.file_assignment_per_instance_[instance_id].push_back(fd.second);
}
} else {
// When not using the deterministic scan range assignment, the scan ranges are
// balanced round robin across fragment instances for the purpose of issuing
// initial scan ranges.
int files_per_instance = file_descs.size() / instance_ctxs.size();
int remainder = file_descs.size() % instance_ctxs.size();
int num_lists = min(file_descs.size(), instance_ctxs.size());
auto fd_it = file_descs.begin();
for (int i = 0; i < num_lists; ++i) {
vector<HdfsFileDesc*>* curr_file_list =
&shared_state_
.file_assignment_per_instance_[instance_ctxs[i]->fragment_instance_id];
for (int j = 0; j < files_per_instance + (i < remainder); ++j) {
curr_file_list->push_back(fd_it->second);
++fd_it;
}
}
DCHECK(fd_it == file_descs.end());
}
return Status::OK();
}
Tuple* HdfsScanPlanNode::InitTemplateTuple(
const std::vector<ScalarExprEvaluator*>& evals, MemPool* pool) const {
if (partition_key_slots_.empty() && !HasVirtualColumnInTemplateTuple()) return nullptr;
Tuple* template_tuple = Tuple::Create(tuple_desc_->byte_size(), pool);
for (int i = 0; i < partition_key_slots_.size(); ++i) {
const SlotDescriptor* slot_desc = partition_key_slots_[i];
ScalarExprEvaluator* eval = evals[slot_desc->col_pos()];
// Exprs guaranteed to be literals, so can safely be evaluated without a row.
RawValue::Write(eval->GetValue(NULL), template_tuple, slot_desc, nullptr);
}
return template_tuple;
}
int HdfsScanPlanNode::GetMaterializedSlotIdx(const std::vector<int>& path) const {
auto result = path_to_materialized_slot_idx_.find(path);
if (result == path_to_materialized_slot_idx_.end()) {
return HdfsScanNodeBase::SKIP_COLUMN;
}
return result->second;
}
void HdfsScanPlanNode::Close() {
TTupleId tuple_id = tnode_->hdfs_scan_node.tuple_id;
for (auto& tid_conjunct : conjuncts_map_) {
// PlanNode::conjuncts_ are already closed in PlanNode::Close()
if (tid_conjunct.first == tuple_id) continue;
ScalarExpr::Close(tid_conjunct.second);
}
ScalarExpr::Close(stats_conjuncts_);
if (shared_state_.template_pool_.get() != nullptr) {
shared_state_.template_pool_->FreeAll();
}
PlanNode::Close();
}
Status HdfsScanPlanNode::CreateExecNode(RuntimeState* state, ExecNode** node) const {
ObjectPool* pool = state->obj_pool();
*node = pool->Add(tnode_->hdfs_scan_node.use_mt_scan_node ?
static_cast<HdfsScanNodeBase*>(
new HdfsScanNodeMt(pool, *this, state->desc_tbl())) :
static_cast<HdfsScanNodeBase*>(
new HdfsScanNode(pool, *this, state->desc_tbl())));
return Status::OK();
}
HdfsScanNodeBase::HdfsScanNodeBase(ObjectPool* pool, const HdfsScanPlanNode& pnode,
const THdfsScanNode& hdfs_scan_node, const DescriptorTbl& descs)
: ScanNode(pool, pnode, descs),
stats_tuple_id_(
hdfs_scan_node.__isset.stats_tuple_id ? hdfs_scan_node.stats_tuple_id : -1),
stats_conjuncts_(pnode.stats_conjuncts_),
stats_tuple_desc_(
stats_tuple_id_ == -1 ? nullptr : descs.GetTupleDescriptor(stats_tuple_id_)),
skip_header_line_count_(hdfs_scan_node.__isset.skip_header_line_count ?
hdfs_scan_node.skip_header_line_count :
0),
tuple_id_(pnode.tuple_id_),
count_star_slot_offset_(hdfs_scan_node.__isset.count_star_slot_offset ?
hdfs_scan_node.count_star_slot_offset :
-1),
is_partition_key_scan_(hdfs_scan_node.is_partition_key_scan),
tuple_desc_(pnode.tuple_desc_),
hdfs_table_(pnode.hdfs_table_),
avro_schema_(*pnode.avro_schema_.get()),
conjuncts_map_(pnode.conjuncts_map_),
thrift_dict_filter_conjuncts_map_(hdfs_scan_node.__isset.dictionary_filter_conjuncts ?
&hdfs_scan_node.dictionary_filter_conjuncts :
nullptr),
codegend_fn_map_(pnode.codegend_fn_map_),
is_materialized_col_(pnode.is_materialized_col_),
materialized_slots_(pnode.materialized_slots_),
partition_key_slots_(pnode.partition_key_slots_),
virtual_column_slots_(pnode.virtual_column_slots_),
disks_accessed_bitmap_(TUnit::UNIT, 0),
active_hdfs_read_thread_counter_(TUnit::UNIT, 0),
shared_state_(const_cast<ScanRangeSharedState*>(&(pnode.shared_state_))),
deterministic_scanrange_assignment_(
hdfs_scan_node.deterministic_scanrange_assignment),
file_metadata_utils_(this) {}
HdfsScanNodeBase::~HdfsScanNodeBase() {}
Status HdfsScanNodeBase::Prepare(RuntimeState* state) {
SCOPED_TIMER(runtime_profile_->total_time_counter());
RETURN_IF_ERROR(ScanNode::Prepare(state));
AddBytesReadCounters();
// Prepare collection conjuncts
for (const auto& entry: conjuncts_map_) {
TupleDescriptor* tuple_desc = state->desc_tbl().GetTupleDescriptor(entry.first);
// conjuncts_ are already prepared in ExecNode::Prepare(), don't try to prepare again
if (tuple_desc == tuple_desc_) {
conjunct_evals_map_[entry.first] = conjunct_evals();
} else {
DCHECK(conjunct_evals_map_[entry.first].empty());
RETURN_IF_ERROR(ScalarExprEvaluator::Create(entry.second, state, pool_,
expr_perm_pool(), expr_results_pool(), &conjunct_evals_map_[entry.first]));
}
}
// Prepare stats statistics conjuncts.
if (stats_tuple_id_ != -1) {
RETURN_IF_ERROR(ScalarExprEvaluator::Create(stats_conjuncts_, state, pool_,
expr_perm_pool(), expr_results_pool(), &stats_conjunct_evals_));
}
// Check if reservation was enough to allocate at least one buffer. The
// reservation calculation in HdfsScanNode.java should guarantee this.
// Hitting this error indicates a misconfiguration or bug.
int64_t min_buffer_size = ExecEnv::GetInstance()->disk_io_mgr()->min_buffer_size();
if (scan_range_params_->size() > 0
&& resource_profile_.min_reservation < min_buffer_size) {
return Status(TErrorCode::INTERNAL_ERROR,
Substitute("HDFS scan min reservation $0 must be >= min buffer size $1",
resource_profile_.min_reservation, min_buffer_size));
}
// One-time initialization of state that is constant across scan ranges
iceberg_partition_filtering_pool_.reset(new MemPool(mem_tracker()));
runtime_profile()->AddInfoString("Table Name", hdfs_table_->fully_qualified_name());
if (HasRowBatchQueue()) {
// Add per volume stats to the runtime profile for Non MT scan node.
PerVolumeStats per_volume_stats;
stringstream str;
UpdateHdfsSplitStats(*scan_range_params_, &per_volume_stats);
PrintHdfsSplitStats(per_volume_stats, &str);
runtime_profile()->AddInfoString(HDFS_SPLIT_STATS_DESC, str.str());
}
return Status::OK();
}
void HdfsScanPlanNode::Codegen(FragmentState* state) {
DCHECK(state->ShouldCodegen());
PlanNode::Codegen(state);
if (IsNodeCodegenDisabled()) return;
for (const THdfsFileFormat::type format : tnode_->hdfs_scan_node.file_formats) {
llvm::Function* fn;
Status status;
switch (format) {
case THdfsFileFormat::TEXT:
status = HdfsTextScanner::Codegen(this, state, &fn);
break;
case THdfsFileFormat::SEQUENCE_FILE:
status = HdfsSequenceScanner::Codegen(this, state, &fn);
break;
case THdfsFileFormat::AVRO:
status = HdfsAvroScanner::Codegen(this, state, &fn);
break;
case THdfsFileFormat::PARQUET:
case THdfsFileFormat::ORC:
status = HdfsColumnarScanner::Codegen(this, state, &fn);
break;
default:
// No codegen for this format
fn = nullptr;
status = Status::Expected("Not implemented for this format.");
}
DCHECK(fn != nullptr || !status.ok());
const char* format_name = _THdfsFileFormat_VALUES_TO_NAMES.find(format)->second;
if (status.ok()) {
LlvmCodeGen* codegen = state->codegen();
DCHECK(codegen != nullptr);
codegen->AddFunctionToJit(
fn, &codegend_fn_map_[static_cast<THdfsFileFormat::type>(format)]);
}
AddCodegenStatus(status, format_name);
}
}
Status HdfsScanNodeBase::Open(RuntimeState* state) {
RETURN_IF_ERROR(ScanNode::Open(state));
// Open collection conjuncts
for (auto& entry: conjunct_evals_map_) {
// conjuncts_ are already opened in ExecNode::Open()
if (entry.first == tuple_id_) continue;
RETURN_IF_ERROR(ScalarExprEvaluator::Open(entry.second, state));
}
// Open stats conjuncts
RETURN_IF_ERROR(ScalarExprEvaluator::Open(stats_conjunct_evals_, state));
RETURN_IF_ERROR(ClaimBufferReservation(state));
reader_context_ = ExecEnv::GetInstance()->disk_io_mgr()->RegisterContext();
// Initialize HdfsScanNode specific counters
hdfs_read_timer_ = PROFILE_TotalRawHdfsReadTime.Instantiate(runtime_profile());
hdfs_open_file_timer_ =
PROFILE_TotalRawHdfsOpenFileTime.Instantiate(runtime_profile());
per_read_thread_throughput_counter_ =
PROFILE_PerReadThreadRawHdfsThroughput.Instantiate(runtime_profile(),
bind<int64_t>(&RuntimeProfile::UnitsPerSecond, bytes_read_counter_,
hdfs_read_timer_));
scan_ranges_complete_counter_ =
PROFILE_ScanRangesComplete.Instantiate(runtime_profile());
collection_items_read_counter_ =
PROFILE_CollectionItemsRead.Instantiate(runtime_profile());
if (DiskInfo::num_disks() < 64) {
num_disks_accessed_counter_ =
PROFILE_NumDisksAccessed.Instantiate(runtime_profile());
} else {
num_disks_accessed_counter_ = NULL;
}
data_cache_hit_count_ = PROFILE_DataCacheHitCount.Instantiate(runtime_profile());
data_cache_partial_hit_count_ =
PROFILE_DataCachePartialHitCount.Instantiate(runtime_profile());
data_cache_miss_count_ = PROFILE_DataCacheMissCount.Instantiate(runtime_profile());
data_cache_hit_bytes_ = PROFILE_DataCacheHitBytes.Instantiate(runtime_profile());
data_cache_miss_bytes_ = PROFILE_DataCacheMissBytes.Instantiate(runtime_profile());
reader_context_->set_bytes_read_counter(bytes_read_counter());
reader_context_->set_read_timer(hdfs_read_timer_);
reader_context_->set_open_file_timer(hdfs_open_file_timer_);
reader_context_->set_active_read_thread_counter(&active_hdfs_read_thread_counter_);
reader_context_->set_disks_accessed_bitmap(&disks_accessed_bitmap_);
reader_context_->set_data_cache_hit_counter(data_cache_hit_count_);
reader_context_->set_data_cache_partial_hit_counter(data_cache_partial_hit_count_);
reader_context_->set_data_cache_miss_counter(data_cache_miss_count_);
reader_context_->set_data_cache_hit_bytes_counter(data_cache_hit_bytes_);
reader_context_->set_data_cache_miss_bytes_counter(data_cache_miss_bytes_);
average_hdfs_read_thread_concurrency_ =
PROFILE_AverageHdfsReadThreadConcurrency.Instantiate(runtime_profile(),
&active_hdfs_read_thread_counter_);
initial_range_ideal_reservation_stats_ =
PROFILE_InitialRangeIdealReservation.Instantiate(runtime_profile());
initial_range_actual_reservation_stats_ =
PROFILE_InitialRangeActualReservation.Instantiate(runtime_profile());
bytes_read_local_ = PROFILE_BytesReadLocal.Instantiate(runtime_profile());
bytes_read_short_circuit_ =
PROFILE_BytesReadShortCircuit.Instantiate(runtime_profile());
bytes_read_dn_cache_ = PROFILE_BytesReadDataNodeCache.Instantiate(runtime_profile());
bytes_read_encrypted_ = PROFILE_BytesReadEncrypted.Instantiate(runtime_profile());
bytes_read_ec_ = PROFILE_BytesReadErasureCoded.Instantiate(runtime_profile());
num_remote_ranges_ = PROFILE_RemoteScanRanges.Instantiate(runtime_profile());
unexpected_remote_bytes_ =
PROFILE_BytesReadRemoteUnexpected.Instantiate(runtime_profile());
cached_file_handles_hit_count_ =
PROFILE_CachedFileHandlesHitCount.Instantiate(runtime_profile());
cached_file_handles_miss_count_ =
PROFILE_CachedFileHandlesMissCount.Instantiate(runtime_profile());
max_compressed_text_file_length_ =
PROFILE_MaxCompressedTextFileLength.Instantiate(runtime_profile());
scanner_io_wait_time_ = PROFILE_ScannerIoWaitTime.Instantiate(runtime_profile());
hdfs_read_thread_concurrency_bucket_ = runtime_profile()->AddBucketingCounters(
&active_hdfs_read_thread_counter_,
ExecEnv::GetInstance()->disk_io_mgr()->num_total_disks() + 1);
counters_running_ = true;
return Status::OK();
}
Status HdfsScanNodeBase::Reset(RuntimeState* state, RowBatch* row_batch) {
DCHECK(false) << "Internal error: Scan nodes should not appear in subplans.";
return Status("Internal error: Scan nodes should not appear in subplans.");
}
void HdfsScanNodeBase::Close(RuntimeState* state) {
if (is_closed()) return;
if (reader_context_ != nullptr) {
// Need to wait for all the active scanner threads to finish to ensure there is no
// more memory tracked by this scan node's mem tracker.
ExecEnv::GetInstance()->disk_io_mgr()->UnregisterContext(reader_context_.get());
}
StopAndFinalizeCounters();
// There should be no active hdfs read threads.
DCHECK_EQ(active_hdfs_read_thread_counter_.value(), 0);
if (iceberg_partition_filtering_pool_.get() != nullptr) {
iceberg_partition_filtering_pool_->FreeAll();
}
// Close collection conjuncts
for (auto& tid_conjunct_eval : conjunct_evals_map_) {
// ExecNode::conjunct_evals_ are already closed in ExecNode::Close()
if (tid_conjunct_eval.first == tuple_id_) continue;
ScalarExprEvaluator::Close(tid_conjunct_eval.second, state);
}
// Close stats conjunct
ScalarExprEvaluator::Close(stats_conjunct_evals_, state);
ScanNode::Close(state);
}
Status HdfsScanNodeBase::IssueInitialScanRanges(RuntimeState* state) {
DCHECK(!initial_ranges_issued_.Load());
initial_ranges_issued_.Store(true);
// We want to decrement this remaining_scan_range_submissions in all cases.
auto remaining_scan_range_submissions_trigger =
MakeScopeExitTrigger([&](){ UpdateRemainingScanRangeSubmissions(-1); });
// No need to issue ranges with limit 0.
if (ReachedLimitShared()) {
DCHECK_EQ(limit_, 0);
return Status::OK();
}
if (filter_ctxs_.size() > 0) WaitForRuntimeFilters();
// Apply dynamic partition-pruning per-file.
HdfsFileDesc::FileFormatsMap matching_per_type_files;
std::vector<HdfsFileDesc*>* file_list =
shared_state_->GetFilesForIssuingScanRangesForInstance(
runtime_state_->instance_ctx().fragment_instance_id);
if (file_list == nullptr) return Status::OK();
for (HdfsFileDesc* file : *file_list) {
if (FilePassesFilterPredicates(state, file, filter_ctxs_)) {
matching_per_type_files[file->file_format].push_back(file);
} else {
SkipFile(file->file_format, file);
}
}
// Issue initial ranges for all file types. Only call functions for file types that
// actually exist - trying to add empty lists of ranges can result in spurious
// CANCELLED errors - see IMPALA-6564.
for (auto& entry : matching_per_type_files) {
if (entry.second.empty()) continue;
// Randomize the order this node processes the files. We want to do this to avoid
// issuing remote reads to the same DN from different impalads. In file formats such
// as avro/seq/rc (i.e. splittable with a header), every node first reads the header.
// If every node goes through the files in the same order, all the remote reads are
// for the same file meaning a few DN serves a lot of remote reads at the same time.
random_shuffle(entry.second.begin(), entry.second.end());
switch (entry.first) {
case THdfsFileFormat::PARQUET:
RETURN_IF_ERROR(HdfsParquetScanner::IssueInitialRanges(this, entry.second));
break;
case THdfsFileFormat::TEXT:
RETURN_IF_ERROR(HdfsTextScanner::IssueInitialRanges(this, entry.second));
break;
case THdfsFileFormat::SEQUENCE_FILE:
case THdfsFileFormat::RC_FILE:
case THdfsFileFormat::AVRO:
RETURN_IF_ERROR(BaseSequenceScanner::IssueInitialRanges(this, entry.second));
break;
case THdfsFileFormat::ORC:
RETURN_IF_ERROR(HdfsOrcScanner::IssueInitialRanges(this, entry.second));
break;
case THdfsFileFormat::JSON:
RETURN_IF_ERROR(HdfsJsonScanner::IssueInitialRanges(this, entry.second));
break;
default:
DCHECK(false) << "Unexpected file type " << entry.first;
}
}
// Except for BaseSequenceScanner, IssueInitialRanges() takes care of
// issuing all the ranges. For BaseSequenceScanner, IssueInitialRanges()
// will have incremented the counter.
return Status::OK();
}
bool HdfsScanNodeBase::FilePassesFilterPredicates(RuntimeState* state, HdfsFileDesc* file,
const vector<FilterContext>& filter_ctxs) {
#ifndef NDEBUG
if (FLAGS_skip_file_runtime_filtering) return true;
#endif
if (filter_ctxs_.size() == 0) return true;
ScanRangeMetadata* metadata =
static_cast<ScanRangeMetadata*>(file->splits[0]->meta_data());
if (hdfs_table_->IsIcebergTable()) {
return IcebergPartitionPassesFilters(
metadata->partition_id, FilterStats::FILES_KEY, filter_ctxs, file, state);
} else {
return PartitionPassesFilters(metadata->partition_id, FilterStats::FILES_KEY,
filter_ctxs);
}
}
void HdfsScanNodeBase::SkipScanRange(io::ScanRange* scan_range) {
// Avoid leaking unread buffers in scan_range.
scan_range->Cancel(Status::CancelledInternal("HDFS partition pruning"));
ScanRangeMetadata* metadata = static_cast<ScanRangeMetadata*>(scan_range->meta_data());
int64_t partition_id = metadata->partition_id;
HdfsPartitionDescriptor* partition = hdfs_table_->GetPartition(partition_id);
DCHECK(partition != nullptr) << "table_id=" << hdfs_table_->id()
<< " partition_id=" << partition_id << "\n"
<< runtime_state_->instance_ctx();
const HdfsFileDesc* desc = GetFileDesc(partition_id, *scan_range->file_string());
if (metadata->is_sequence_header) {
// File ranges haven't been issued yet, skip entire file.
UpdateRemainingScanRangeSubmissions(-1);
SkipFile(partition->file_format(), desc);
} else {
// Mark this scan range as done.
HdfsScanNodeBase::RangeComplete(
partition->file_format(), desc->file_compression, true);
}
}
Status HdfsScanNodeBase::StartNextScanRange(const std::vector<FilterContext>& filter_ctxs,
int64_t* reservation, ScanRange** scan_range) {
DiskIoMgr* io_mgr = ExecEnv::GetInstance()->disk_io_mgr();
bool needs_buffers;
// Loop until we've got a scan range or run out of ranges.
do {
RETURN_IF_ERROR(GetNextScanRangeToRead(scan_range, &needs_buffers));
if (*scan_range == nullptr) return Status::OK();
if (filter_ctxs.size() > 0) {
int64_t partition_id =
static_cast<ScanRangeMetadata*>((*scan_range)->meta_data())->partition_id;
if (!hdfs_table()->IsIcebergTable() &&
!PartitionPassesFilters(partition_id, FilterStats::SPLITS_KEY, filter_ctxs)) {
SkipScanRange(*scan_range);
*scan_range = nullptr;
}
}
} while (*scan_range == nullptr);
if (needs_buffers) {
// Check if we should increase our reservation to read this range more efficiently.
// E.g. if we are scanning a large text file, we might want extra I/O buffers to
// improve throughput. Note that if this is a columnar format like Parquet,
// '*scan_range' is the small footer range only so we won't request an increase.
int64_t ideal_scan_range_reservation =
io_mgr->ComputeIdealBufferReservation((*scan_range)->bytes_to_read());
*reservation = IncreaseReservationIncrementally(*reservation, ideal_scan_range_reservation);
initial_range_ideal_reservation_stats_->UpdateCounter(ideal_scan_range_reservation);
initial_range_actual_reservation_stats_->UpdateCounter(*reservation);
RETURN_IF_ERROR(
io_mgr->AllocateBuffersForRange(buffer_pool_client(), *scan_range, *reservation));
}
return Status::OK();
}
int64_t HdfsScanNodeBase::IncreaseReservationIncrementally(int64_t curr_reservation,
int64_t ideal_reservation) {
DiskIoMgr* io_mgr = ExecEnv::GetInstance()->disk_io_mgr();
// Check if we could use at least one more max-sized I/O buffer for this range. Don't
// increase in smaller increments since we may not be able to use additional smaller
// buffers.
while (curr_reservation < ideal_reservation) {
// Increase to the next I/O buffer multiple or to the ideal reservation.
int64_t target = min(ideal_reservation,
BitUtil::RoundUpToPowerOf2(curr_reservation + 1, io_mgr->max_buffer_size()));
DCHECK_LT(curr_reservation, target);
bool increased = buffer_pool_client()->IncreaseReservation(target - curr_reservation);
VLOG_FILE << "Increasing reservation from "
<< PrettyPrinter::PrintBytes(curr_reservation) << " to "
<< PrettyPrinter::PrintBytes(target) << " "
<< (increased ? "succeeded" : "failed");
if (!increased) break;
curr_reservation = target;
}
return curr_reservation;
}
ScanRange* HdfsScanNodeBase::AllocateScanRange(const ScanRange::FileInfo &fi,
int64_t len, int64_t offset, int64_t partition_id, int disk_id, bool expected_local,
const BufferOpts& buffer_opts, const ScanRange* original_split) {
ScanRangeMetadata* metadata =
shared_state_->obj_pool()->Add(new ScanRangeMetadata(partition_id, original_split));
return AllocateScanRange(fi, len, offset, {}, metadata, disk_id, expected_local,
buffer_opts);
}
ScanRange* HdfsScanNodeBase::AllocateScanRange(const ScanRange::FileInfo &fi,
int64_t len, int64_t offset, vector<ScanRange::SubRange>&& sub_ranges,
int64_t partition_id, int disk_id, bool expected_local,
const BufferOpts& buffer_opts, const ScanRange* original_split) {
ScanRangeMetadata* metadata =
shared_state_->obj_pool()->Add(new ScanRangeMetadata(partition_id, original_split));
return AllocateScanRange(fi, len, offset, move(sub_ranges), metadata,
disk_id, expected_local, buffer_opts);
}
ScanRange* HdfsScanNodeBase::AllocateScanRange(const ScanRange::FileInfo &fi,
int64_t len, int64_t offset, vector<ScanRange::SubRange>&& sub_ranges,
ScanRangeMetadata* metadata, int disk_id, bool expected_local,
const BufferOpts& buffer_opts) {
// Require that the scan range is within [0, file_length). While this cannot be used
// to guarantee safety (file_length metadata may be stale), it avoids different
// behavior between Hadoop FileSystems (e.g. s3n hdfsSeek() returns error when seeking
// beyond the end of the file).
DCHECK_LE(offset + len, GetFileDesc(metadata->partition_id, fi.filename)->file_length)
<< "Scan range beyond end of file (offset=" << offset << ", len=" << len << ")";
return ScanRange::AllocateScanRange(shared_state_->obj_pool(), fi, len, offset,
move(sub_ranges), metadata, disk_id, expected_local, buffer_opts);
}
const CodegenFnPtrBase* HdfsScanNodeBase::GetCodegenFn(THdfsFileFormat::type type) {
HdfsScanPlanNode::CodegendFnMap::const_iterator it = codegend_fn_map_.find(type);
if (it == codegend_fn_map_.end()) return nullptr;
return &it->second;
}
Status HdfsScanNodeBase::CreateAndOpenScannerHelper(
ScannerContext* context, scoped_ptr<HdfsScanner>* scanner) {
DCHECK(context != nullptr);
DCHECK(scanner->get() == nullptr);
const HdfsFileDesc* file_desc = context->GetStream(0)->file_desc();
THdfsCompression::type compression = file_desc->file_compression;
switch (file_desc->file_format) {
case THdfsFileFormat::TEXT:
if (HdfsTextScanner::HasBuiltinSupport(compression)) {
scanner->reset(new HdfsTextScanner(this, runtime_state_));
} else {
// No builtin support - we must have loaded the plugin in IssueInitialRanges().
auto it = _THdfsCompression_VALUES_TO_NAMES.find(compression);
DCHECK(it != _THdfsCompression_VALUES_TO_NAMES.end())
<< "Already issued ranges for this compression type.";
scanner->reset(HdfsPluginTextScanner::GetHdfsPluginTextScanner(
this, runtime_state_, it->second));
}
break;
case THdfsFileFormat::SEQUENCE_FILE:
scanner->reset(new HdfsSequenceScanner(this, runtime_state_));
break;
case THdfsFileFormat::RC_FILE:
scanner->reset(new HdfsRCFileScanner(this, runtime_state_));
break;
case THdfsFileFormat::AVRO:
scanner->reset(new HdfsAvroScanner(this, runtime_state_));
break;
case THdfsFileFormat::PARQUET:
scanner->reset(new HdfsParquetScanner(this, runtime_state_));
break;
case THdfsFileFormat::ORC:
scanner->reset(new HdfsOrcScanner(this, runtime_state_));
break;
case THdfsFileFormat::JSON:
if (HdfsJsonScanner::HasBuiltinSupport(compression)) {
scanner->reset(new HdfsJsonScanner(this, runtime_state_));
} else {
return Status("Scanning compressed Json file is not implemented yet.");
}
break;
default:
return Status(
Substitute("Unknown Hdfs file format type: $0", file_desc->file_format));
}
DCHECK(scanner->get() != nullptr);
RETURN_IF_ERROR(scanner->get()->Open(context));
// Inject the error after the scanner is opened, to test the scanner close path.
return ScanNodeDebugAction(TExecNodePhase::PREPARE_SCANNER);
}
void HdfsScanNodeBase::InitNullCollectionValues(const TupleDescriptor* tuple_desc,
Tuple* tuple) const {
for (const SlotDescriptor* slot_desc: tuple_desc->collection_slots()) {
CollectionValue* slot = reinterpret_cast<CollectionValue*>(
tuple->GetSlot(slot_desc->tuple_offset()));
if (tuple->IsNull(slot_desc->null_indicator_offset())) {
*slot = CollectionValue();
continue;
}
// Recursively traverse collection items.
const TupleDescriptor* item_desc = slot_desc->children_tuple_descriptor();
if (item_desc->collection_slots().empty()) continue;
for (int i = 0; i < slot->num_tuples; ++i) {
int item_offset = i * item_desc->byte_size();
Tuple* collection_item = reinterpret_cast<Tuple*>(slot->ptr + item_offset);
InitNullCollectionValues(item_desc, collection_item);
}
}
}
void HdfsScanNodeBase::InitNullCollectionValues(RowBatch* row_batch) const {
DCHECK_EQ(row_batch->row_desc()->tuple_descriptors().size(), 1);
const TupleDescriptor& tuple_desc =
*row_batch->row_desc()->tuple_descriptors()[0];
if (tuple_desc.collection_slots().empty()) return;
for (int i = 0; i < row_batch->num_rows(); ++i) {
Tuple* tuple = row_batch->GetRow(i)->GetTuple(0);
DCHECK(tuple != NULL);
InitNullCollectionValues(&tuple_desc, tuple);
}
}
bool HdfsScanNodeBase::PartitionPassesFilters(int32_t partition_id,
const string& stats_name, const vector<FilterContext>& filter_ctxs) {
DCHECK(!hdfs_table()->IsIcebergTable());
if (filter_ctxs.empty()) return true;
if (FilterContext::CheckForAlwaysFalse(stats_name, filter_ctxs)) return false;
DCHECK_EQ(filter_ctxs.size(), filter_ctxs_.size())
<< "Mismatched number of filter contexts";
Tuple* template_tuple = GetTemplateTupleForPartitionId(partition_id);
// Defensive - if template_tuple is NULL, there can be no filters on partition columns.
if (template_tuple == nullptr) return true;
TupleRow* tuple_row_mem = reinterpret_cast<TupleRow*>(&template_tuple);
for (const FilterContext& ctx: filter_ctxs) {
if (!ctx.filter->IsBoundByPartitionColumn(id_)) {
continue;
}
bool has_filter = ctx.filter->HasFilter();
bool passed_filter = !has_filter || ctx.Eval(tuple_row_mem);
ctx.stats->IncrCounters(stats_name, 1, has_filter, !passed_filter);
if (!passed_filter) return false;
}
return true;
}
bool HdfsScanNodeBase::IcebergPartitionPassesFilters(int64_t partition_id,
const string& stats_name, const vector<FilterContext>& filter_ctxs,
HdfsFileDesc* file, RuntimeState* state) {
DCHECK(hdfs_table()->IsIcebergTable());
file_metadata_utils_.SetFile(state, file);
// Create the template tuple based on file metadata
std::map<const SlotId, const SlotDescriptor*> slot_descs_written;
Tuple* template_tuple = file_metadata_utils_.CreateTemplateTuple(partition_id,
iceberg_partition_filtering_pool_.get(), &slot_descs_written);
// Defensive - if template_tuple is NULL, there can be no filters on partition columns.
if (template_tuple == nullptr) return true;
// Try to evaluate all filter_ctxs on template_tuple
for (const FilterContext& ctx: filter_ctxs) {
bool skip_filter = false;
const auto& expr = ctx.expr_eval->root();
// Match written SlotDescriptors to SlotDescriptors available in this FilterContext
vector<SlotId> slot_ids;
expr.GetSlotIds(&slot_ids);
for (SlotId filter_slot_id : slot_ids) {
if (slot_descs_written.find(filter_slot_id) == slot_descs_written.end()) {
skip_filter = true;
break;
}
}
// Do not evaluate this filter when the slots are not present in the template_tuple
if (skip_filter) continue;
// Evaluate filter on template_tuple
TupleRow* row = reinterpret_cast<TupleRow*>(&template_tuple);
bool has_filter = ctx.filter->HasFilter();
bool passed_filter = !has_filter || ctx.Eval(row);
ctx.stats->IncrCounters(stats_name, 1, has_filter, !passed_filter);
if (!passed_filter) return false;
}
iceberg_partition_filtering_pool_->FreeAll();
return true;
}
void HdfsScanNodeBase::RangeComplete(const THdfsFileFormat::type& file_type,
const THdfsCompression::type& compression_type, bool skipped) {
vector<THdfsCompression::type> types;
types.push_back(compression_type);
RangeComplete(file_type, types, skipped);
}
void HdfsScanNodeBase::RangeComplete(const THdfsFileFormat::type& file_type,
const vector<THdfsCompression::type>& compression_types, bool skipped) {
scan_ranges_complete_counter_->Add(1);
shared_state_->progress().Update(1);
HdfsCompressionTypesSet compression_set;
for (int i = 0; i < compression_types.size(); ++i) {
compression_set.AddType(compression_types[i]);
}
++file_type_counts_[std::make_tuple(file_type, skipped, compression_set)];
}
void HdfsScanNodeBase::SkipFile(const THdfsFileFormat::type& file_type,
const HdfsFileDesc* file) {
for (int i = 0; i < file->splits.size(); ++i) {
RangeComplete(file_type, file->file_compression, true);
}
}
void HdfsScanPlanNode::ComputeSlotMaterializationOrder(
const DescriptorTbl& desc_tbl, vector<int>* order) const {
const vector<ScalarExpr*>& conjuncts = conjuncts_;
// Initialize all order to be conjuncts.size() (after the last conjunct)
order->insert(order->begin(), materialized_slots_.size(), conjuncts.size());
vector<SlotId> slot_ids;
for (int conjunct_idx = 0; conjunct_idx < conjuncts.size(); ++conjunct_idx) {
slot_ids.clear();
int num_slots = conjuncts[conjunct_idx]->GetSlotIds(&slot_ids);
for (int j = 0; j < num_slots; ++j) {
SlotDescriptor* slot_desc = desc_tbl.GetSlotDescriptor(slot_ids[j]);
int slot_idx = GetMaterializedSlotIdx(slot_desc->col_path());
// slot_idx == -1 means this was a partition key slot which is always
// materialized before any slots.
if (slot_idx == -1) continue;
// If this slot hasn't been assigned an order, assign it be materialized
// before evaluating conjuncts[i]
if ((*order)[slot_idx] == conjuncts.size()) {
(*order)[slot_idx] = conjunct_idx;
}
}
}
}
bool HdfsScanPlanNode::HasVirtualColumnInTemplateTuple() const {
for (SlotDescriptor* sd : virtual_column_slots_) {
DCHECK(sd->IsVirtual());
if (sd->virtual_column_type() == TVirtualColumnType::FILE_POSITION) {
// We return false at the end of the function if there are no virtual
// columns in the template tuple.
continue;
} else if (sd->virtual_column_type() == TVirtualColumnType::INPUT_FILE_NAME ||
sd->virtual_column_type() == TVirtualColumnType::PARTITION_SPEC_ID ||
sd->virtual_column_type() == TVirtualColumnType::ICEBERG_PARTITION_SERIALIZED ||
sd->virtual_column_type() == TVirtualColumnType::ICEBERG_DATA_SEQUENCE_NUMBER ||
sd->virtual_column_type() == TVirtualColumnType::ICEBERG_FIRST_ROW_ID) {
return true;
} else {
// Adding DCHECK here so we don't forget to update this when adding new virtual
// column.
DCHECK(false);
}
}
return false;
}
void HdfsScanNodeBase::TransferToSharedStatePool(MemPool* pool) {
shared_state_->TransferToSharedStatePool(pool);
}
void HdfsScanNodeBase::UpdateHdfsSplitStats(
const google::protobuf::RepeatedPtrField<ScanRangeParamsPB>& scan_range_params_list,
PerVolumeStats* per_volume_stats) {
pair<int, int64_t> init_value(0, 0);
for (const ScanRangeParamsPB& scan_range_params : scan_range_params_list) {
const ScanRangePB& scan_range = scan_range_params.scan_range();
if (!scan_range.has_hdfs_file_split()) continue;
const HdfsFileSplitPB& split = scan_range.hdfs_file_split();
pair<int, int64_t>* stats =
FindOrInsert(per_volume_stats, scan_range_params.volume_id(), init_value);
++(stats->first);
stats->second += split.length();
}
}
void HdfsScanNodeBase::PrintHdfsSplitStats(const PerVolumeStats& per_volume_stats,
stringstream* ss) {
for (PerVolumeStats::const_iterator i = per_volume_stats.begin();
i != per_volume_stats.end(); ++i) {
(*ss) << i->first << ":" << i->second.first << "/"
<< PrettyPrinter::Print(i->second.second, TUnit::BYTES) << " ";
}
}
void HdfsScanNodeBase::StopAndFinalizeCounters() {
if (!counters_running_) return;
counters_running_ = false;
runtime_profile_->StopPeriodicCounters();
// Output hdfs read thread concurrency into info string
stringstream ss;
for (int i = 0; i < hdfs_read_thread_concurrency_bucket_->size(); ++i) {
ss << i << ":" << setprecision(4)
<< (*hdfs_read_thread_concurrency_bucket_)[i]->double_value() << "% ";
}
runtime_profile_->AddInfoString("Hdfs Read Thread Concurrency Bucket", ss.str());
// Convert disk access bitmap to num of disk accessed
uint64_t num_disk_bitmap = disks_accessed_bitmap_.value();
int64_t num_disk_accessed = BitUtil::Popcount(num_disk_bitmap);
if (num_disks_accessed_counter_ != NULL) {
num_disks_accessed_counter_->Set(num_disk_accessed);
}
// output completed file types and counts to info string
if (!file_type_counts_.empty()) {
stringstream ss;
{
for (FileTypeCountsMap::const_iterator it = file_type_counts_.begin();
it != file_type_counts_.end(); ++it) {
THdfsFileFormat::type file_format = std::get<0>(it->first);
bool skipped = std::get<1>(it->first);
HdfsCompressionTypesSet compressions_set = std::get<2>(it->first);
int file_cnt = it->second;
if (skipped) {
if (file_format == THdfsFileFormat::PARQUET) {
// If a scan range stored as parquet is skipped, its compression type
// cannot be figured out without reading the data.
ss << file_format << "/" << "Unknown" << "(Skipped):"
<< file_cnt << " ";
} else {
ss << file_format << "/"
<< compressions_set.GetFirstType() << "(Skipped):"
<< file_cnt << " ";
}
} else if (compressions_set.Size() == 1) {
ss << file_format << "/"
<< compressions_set.GetFirstType() << ":" << file_cnt
<< " ";
} else {
ss << file_format << "/" << "(";
bool first = true;
for (auto& elem : _THdfsCompression_VALUES_TO_NAMES) {
THdfsCompression::type type = static_cast<THdfsCompression::type>(
elem.first);
if (!compressions_set.HasType(type)) continue;
if (!first) ss << ",";
ss << type;
first = false;
}
ss << "):" << file_cnt << " ";
}
}
}
runtime_profile_->AddInfoString("File Formats", ss.str());
}
// Output fraction of scanners with codegen enabled
int num_enabled = num_scanners_codegen_enabled_.Load();
int total = num_enabled + num_scanners_codegen_disabled_.Load();
runtime_profile()->AppendExecOption(
Substitute("Codegen enabled: $0 out of $1", num_enabled, total));
// Locking here should not be necessary since bytes_read_per_col_ is only updated inside
// column readers, and all column readers should have completed at this point; however,
// we acquire a read lock in case the update semantics of bytes_read_per_col_ change
{
shared_lock<shared_mutex> bytes_read_per_col_guard_read_lock(
bytes_read_per_col_lock_);
if (!bytes_read_per_col_.empty()) {
auto uncompressed_bytes_counter =
PROFILE_ParquetUncompressedBytesReadPerColumn.Instantiate(runtime_profile());
auto compressed_bytes_counter =
PROFILE_ParquetCompressedBytesReadPerColumn.Instantiate(runtime_profile());
for (const auto& bytes_read : bytes_read_per_col_) {
int64_t uncompressed_bytes_read =
bytes_read.second.uncompressed_bytes_read.Load();
if (uncompressed_bytes_read > 0) {
uncompressed_bytes_counter->UpdateCounter(uncompressed_bytes_read);
}
int64_t compressed_bytes_read = bytes_read.second.compressed_bytes_read.Load();
if (compressed_bytes_read > 0) {
compressed_bytes_counter->UpdateCounter(compressed_bytes_read);
}
}
}
}
if (reader_context_ != nullptr) {
bytes_read_local_->Set(reader_context_->bytes_read_local());
bytes_read_short_circuit_->Set(reader_context_->bytes_read_short_circuit());
bytes_read_dn_cache_->Set(reader_context_->bytes_read_dn_cache());
bytes_read_encrypted_->Set(reader_context_->bytes_read_encrypted());
bytes_read_ec_->Set(reader_context_->bytes_read_ec());
num_remote_ranges_->Set(reader_context_->num_remote_ranges());
unexpected_remote_bytes_->Set(reader_context_->unexpected_remote_bytes());
cached_file_handles_hit_count_->Set(reader_context_->cached_file_handles_hit_count());
cached_file_handles_miss_count_->Set(
reader_context_->cached_file_handles_miss_count());
if (unexpected_remote_bytes_->value() >= UNEXPECTED_REMOTE_BYTES_WARN_THRESHOLD) {
runtime_state_->LogError(ErrorMsg(TErrorCode::GENERAL, Substitute(
"Read $0 of data across network that was expected to be local. Block locality "
"metadata for table '$1.$2' may be stale. This only affects query performance "
"and not result correctness. One of the common causes for this warning is HDFS "
"rebalancer moving some of the file's blocks. If the issue persists, consider "
"running \"INVALIDATE METADATA `$1`.`$2`\".",
PrettyPrinter::Print(unexpected_remote_bytes_->value(), TUnit::BYTES),
hdfs_table_->database(), hdfs_table_->name())));
}
ImpaladMetrics::IO_MGR_BYTES_READ->Increment(bytes_read_counter()->value());
ImpaladMetrics::IO_MGR_LOCAL_BYTES_READ->Increment(
bytes_read_local_->value());
ImpaladMetrics::IO_MGR_SHORT_CIRCUIT_BYTES_READ->Increment(
bytes_read_short_circuit_->value());
ImpaladMetrics::IO_MGR_CACHED_BYTES_READ->Increment(
bytes_read_dn_cache_->value());
ImpaladMetrics::IO_MGR_ENCRYPTED_BYTES_READ->Increment(
bytes_read_encrypted_->value());
ImpaladMetrics::IO_MGR_ERASURE_CODED_BYTES_READ->Increment(
bytes_read_ec_->value());
}
}
Status HdfsScanNodeBase::ScanNodeDebugAction(TExecNodePhase::type phase) {
return ExecDebugAction(phase, runtime_state_);
}
void HdfsScanNodeBase::UpdateBytesRead(
SlotId slot_id, int64_t uncompressed_bytes_read, int64_t compressed_bytes_read) {
// Acquire a read lock first and check if the slot_id is in bytes_read_per_col_, if it
// is then update the value and release the read lock; if not then release the read
// lock, acquire the write lock, and then initialize the slot_id with the give value for
// bytes_read
shared_lock<shared_mutex> bytes_read_per_col_guard_read_lock(
bytes_read_per_col_lock_);
auto bytes_read_itr = bytes_read_per_col_.find(slot_id);
if (bytes_read_itr != bytes_read_per_col_.end()) {
bytes_read_itr->second.uncompressed_bytes_read.Add(uncompressed_bytes_read);
bytes_read_itr->second.compressed_bytes_read.Add(compressed_bytes_read);
} else {
bytes_read_per_col_guard_read_lock.unlock();
lock_guard<shared_mutex> bytes_read_per_col_guard_write_lock(
bytes_read_per_col_lock_);
bytes_read_per_col_[slot_id].uncompressed_bytes_read.Add(uncompressed_bytes_read);
bytes_read_per_col_[slot_id].compressed_bytes_read.Add(compressed_bytes_read);
}
}
const HdfsFileDesc* ScanRangeSharedState::GetFileDesc(
int64_t partition_id, const std::string& filename) {
auto file_desc_map_key = make_pair(partition_id, filename);
DCHECK(file_descs_.find(file_desc_map_key) != file_descs_.end());
return file_descs_[file_desc_map_key];
}
void ScanRangeSharedState::SetFileMetadata(
int64_t partition_id, const string& filename, void* metadata) {
unique_lock<mutex> l(metadata_lock_);
auto file_metadata_map_key = make_pair(partition_id, filename);
DCHECK(per_file_metadata_.find(file_metadata_map_key) == per_file_metadata_.end());
per_file_metadata_[file_metadata_map_key] = metadata;
}
void* ScanRangeSharedState::GetFileMetadata(
int64_t partition_id, const string& filename) {
unique_lock<mutex> l(metadata_lock_);
auto file_metadata_map_key = make_pair(partition_id, filename);
auto it = per_file_metadata_.find(file_metadata_map_key);
if (it == per_file_metadata_.end()) return NULL;
return it->second;
}
Tuple* ScanRangeSharedState::GetTemplateTupleForPartitionId(int64_t partition_id) {
DCHECK(partition_template_tuple_map_.find(partition_id)
!= partition_template_tuple_map_.end());
return partition_template_tuple_map_[partition_id];
}
void ScanRangeSharedState::TransferToSharedStatePool(MemPool* pool) {
unique_lock<mutex> l(metadata_lock_);
template_pool_->AcquireData(pool, false);
}
void ScanRangeSharedState::UpdateRemainingScanRangeSubmissions(int32_t delta) {
int new_val = remaining_scan_range_submissions_.Add(delta);
DCHECK_GE(new_val, 0);
if (!use_mt_scan_node_) return;
if (new_val == 0) {
// Last thread has added its ranges. Acquire lock so that no waiting thread misses the
// last notify.
std::unique_lock<std::mutex> l(scan_range_submission_lock_);
}
range_submission_cv_.NotifyAll();
}
void ScanRangeSharedState::EnqueueScanRange(
const vector<ScanRange*>& ranges, bool at_front) {
DCHECK(use_mt_scan_node_) << "Should only be called by MT scan nodes";
scan_range_queue_.EnqueueRanges(ranges, at_front);
}
Status ScanRangeSharedState::GetNextScanRange(
RuntimeState* state, ScanRange** scan_range) {
DCHECK(use_mt_scan_node_) << "Should only be called by MT scan nodes";
while (true) {
*scan_range = scan_range_queue_.Dequeue();
if (*scan_range != nullptr) return Status::OK();
{
unique_lock<mutex> l(scan_range_submission_lock_);
while (scan_range_queue_.Empty() && remaining_scan_range_submissions_.Load() > 0
&& !state->is_cancelled()) {
range_submission_cv_.Wait(l);
}
}
// No more work to do.
if (scan_range_queue_.Empty() && remaining_scan_range_submissions_.Load() == 0) {
break;
}
if (state->is_cancelled()) return Status::CANCELLED;
}
return Status::OK();
}
void ScanRangeSharedState::AddCancellationHook(RuntimeState* state) {
DCHECK(use_mt_scan_node_) << "Should only be called by MT scan nodes";
state->AddCancellationCV(&scan_range_submission_lock_, &range_submission_cv_);
}
}