<?xml version="1.0" encoding="UTF-8"?>
<!--
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.
-->
<!DOCTYPE concept PUBLIC "-//OASIS//DTD DITA Concept//EN" "concept.dtd">
<concept id="explain_plan">

  <title>Understanding Impala Query Performance - EXPLAIN Plans and Query Profiles</title>
  <titlealts audience="PDF"><navtitle>EXPLAIN Plans and Query Profiles</navtitle></titlealts>
  <prolog>
    <metadata>
      <data name="Category" value="Performance"/>
      <data name="Category" value="Impala"/>
      <data name="Category" value="Querying"/>
      <data name="Category" value="Troubleshooting"/>
      <data name="Category" value="Reports"/>
      <data name="Category" value="Concepts"/>
      <data name="Category" value="Administrators"/>
      <data name="Category" value="Developers"/>
      <data name="Category" value="Data Analysts"/>
    </metadata>
  </prolog>

  <conbody>

    <p>
      To understand the high-level performance considerations for Impala queries, read the output of the
      <codeph>EXPLAIN</codeph> statement for the query. You can get the <codeph>EXPLAIN</codeph> plan without
      actually running the query itself.
    </p>

    <p rev="1.4.0">
      For an overview of the physical performance characteristics for a query, issue the <codeph>SUMMARY</codeph>
      statement in <cmdname>impala-shell</cmdname> immediately after executing a query. This condensed information
      shows which phases of execution took the most time, and how the estimates for memory usage and number of rows
      at each phase compare to the actual values.
    </p>

    <p>
      To understand the detailed performance characteristics for a query, issue the <codeph>PROFILE</codeph>
      statement in <cmdname>impala-shell</cmdname> immediately after executing a query. This low-level information
      includes physical details about memory, CPU, I/O, and network usage, and thus is only available after the
      query is actually run.
    </p>

    <p outputclass="toc inpage"/>

    <p>
      Also, see <xref href="impala_hbase.xml#hbase_performance"/>
      and <xref href="impala_s3.xml#s3_performance"/>
      for examples of interpreting
      <codeph>EXPLAIN</codeph> plans for queries against HBase tables
      <ph rev="2.2.0">and data stored in the Amazon Simple Storage System (S3)</ph>.
    </p>
  </conbody>

  <concept id="perf_explain">

    <title>Using the EXPLAIN Plan for Performance Tuning</title>

    <conbody>

      <p>
        The <codeph><xref href="impala_explain.xml#explain">EXPLAIN</xref></codeph> statement gives you an outline
        of the logical steps that a query will perform, such as how the work will be distributed among the nodes
        and how intermediate results will be combined to produce the final result set. You can see these details
        before actually running the query. You can use this information to check that the query will not operate in
        some very unexpected or inefficient way.
      </p>

<!-- Turn into a conref in ciiu_langref too. Relocate to common.xml. -->

<codeblock conref="impala_explain.xml#explain/explain_plan_simple"/>

      <p conref="../shared/impala_common.xml#common/explain_interpret"/>

      <p>
        The <codeph>EXPLAIN</codeph> plan is also printed at the beginning of the query profile report described in
        <xref href="#perf_profile"/>, for convenience in examining both the logical and physical aspects of the
        query side-by-side.
      </p>

      <p rev="1.2">
        The amount of detail displayed in the <codeph>EXPLAIN</codeph> output is controlled by the
        <xref href="impala_explain_level.xml#explain_level">EXPLAIN_LEVEL</xref> query option. You typically
        increase this setting from <codeph>standard</codeph> to <codeph>extended</codeph> (or from <codeph>1</codeph>
        to <codeph>2</codeph>) when doublechecking the presence of table and column statistics during performance
        tuning, or when estimating query resource usage in conjunction with the resource management features.
      </p>

      <!-- To do:
        This is a good place to have a few examples.
      -->
    </conbody>
  </concept>

  <concept id="perf_summary">

    <title>Using the SUMMARY Report for Performance Tuning</title>

    <conbody>

      <p>
        The <codeph><xref href="impala_shell_commands.xml#shell_commands">SUMMARY</xref></codeph> command within
        the <cmdname>impala-shell</cmdname> interpreter gives you an easy-to-digest overview of the timings for the
        different phases of execution for a query. Like the <codeph>EXPLAIN</codeph> plan, it is easy to see
        potential performance bottlenecks. Like the <codeph>PROFILE</codeph> output, it is available after the
        query is run and so displays actual timing numbers.
      </p>

      <p>
        The <codeph>SUMMARY</codeph> report is also printed at the beginning of the query profile report described
        in <xref href="#perf_profile"/>, for convenience in examining high-level and low-level aspects of the query
        side-by-side.
      </p>

      <p>
        For example, here is a query involving an aggregate function, on a single-node VM. The different stages of
        the query and their timings are shown (rolled up for all nodes), along with estimated and actual values
        used in planning the query. In this case, the <codeph>AVG()</codeph> function is computed for a subset of
        data on each node (stage 01) and then the aggregated results from all nodes are combined at the end (stage
        03). You can see which stages took the most time, and whether any estimates were substantially different
        than the actual data distribution. (When examining the time values, be sure to consider the suffixes such
        as <codeph>us</codeph> for microseconds and <codeph>ms</codeph> for milliseconds, rather than just looking
        for the largest numbers.)
      </p>

<codeblock>[localhost:21000] &gt; select avg(ss_sales_price) from store_sales where ss_coupon_amt = 0;
+---------------------+
| avg(ss_sales_price) |
+---------------------+
| 37.80770926328327   |
+---------------------+
[localhost:21000] &gt; summary;
+--------------+--------+----------+----------+-------+------------+----------+---------------+-----------------+
| Operator     | #Hosts | Avg Time | Max Time | #Rows | Est. #Rows | Peak Mem | Est. Peak Mem | Detail          |
+--------------+--------+----------+----------+-------+------------+----------+---------------+-----------------+
| 03:AGGREGATE | 1      | 1.03ms   | 1.03ms   | 1     | 1          | 48.00 KB | -1 B          | MERGE FINALIZE  |
| 02:EXCHANGE  | 1      | 0ns      | 0ns      | 1     | 1          | 0 B      | -1 B          | UNPARTITIONED   |
| 01:AGGREGATE | 1      | 30.79ms  | 30.79ms  | 1     | 1          | 80.00 KB | 10.00 MB      |                 |
| 00:SCAN HDFS | 1      | 5.45s    | 5.45s    | 2.21M | -1         | 64.05 MB | 432.00 MB     | tpc.store_sales |
+--------------+--------+----------+----------+-------+------------+----------+---------------+-----------------+
</codeblock>

      <p>
        Notice how the longest initial phase of the query is measured in seconds (s), while later phases working on
        smaller intermediate results are measured in milliseconds (ms) or even nanoseconds (ns).
      </p>

      <p>
        Here is an example from a more complicated query, as it would appear in the <codeph>PROFILE</codeph>
        output:
      </p>

<codeblock>Operator              #Hosts   Avg Time   Max Time    #Rows  Est. #Rows  Peak Mem  Est. Peak Mem  Detail
------------------------------------------------------------------------------------------------------------------------
09:MERGING-EXCHANGE        1   79.738us   79.738us        5           5         0        -1.00 B  UNPARTITIONED
05:TOP-N                   3   84.693us   88.810us        5           5  12.00 KB       120.00 B
04:AGGREGATE               3    5.263ms    6.432ms        5           5  44.00 KB       10.00 MB  MERGE FINALIZE
08:AGGREGATE               3   16.659ms   27.444ms   52.52K     600.12K   3.20 MB       15.11 MB  MERGE
07:EXCHANGE                3    2.644ms      5.1ms   52.52K     600.12K         0              0  HASH(o_orderpriority)
03:AGGREGATE               3  342.913ms  966.291ms   52.52K     600.12K  10.80 MB       15.11 MB
02:HASH JOIN               3    2s165ms    2s171ms  144.87K     600.12K  13.63 MB      941.01 KB  INNER JOIN, BROADCAST
|--06:EXCHANGE             3    8.296ms    8.692ms   57.22K      15.00K         0              0  BROADCAST
|  01:SCAN HDFS            2    1s412ms    1s978ms   57.22K      15.00K  24.21 MB      176.00 MB  tpch.orders o
00:SCAN HDFS               3    8s032ms    8s558ms    3.79M     600.12K  32.29 MB      264.00 MB  tpch.lineitem l
</codeblock>
    </conbody>
  </concept>

  <concept id="perf_profile">

    <title>Using the Query Profile for Performance Tuning</title>

    <conbody>

      <p> The <codeph>PROFILE</codeph> command, available in the
          <cmdname>impala-shell</cmdname> interpreter, produces a detailed
        low-level report showing how the most recent query was executed. Unlike
        the <codeph>EXPLAIN</codeph> plan described in <xref
          href="#perf_explain"/>, this information is only available after the
        query has finished. It shows physical details such as the number of
        bytes read, maximum memory usage, and so on for each node. You can use
        this information to determine if the query is I/O-bound or CPU-bound,
        whether some network condition is imposing a bottleneck, whether a
        slowdown is affecting some nodes but not others, and to check that
        recommended configuration settings such as short-circuit local reads are
        in effect. </p>

      <p rev="">
        By default, time values in the profile output reflect the wall-clock time taken by an operation.
        For values denoting system time or user time, the measurement unit is reflected in the metric
        name, such as <codeph>ScannerThreadsSysTime</codeph> or <codeph>ScannerThreadsUserTime</codeph>.
        For example, a multi-threaded I/O operation might show a small figure for wall-clock time,
        while the corresponding system time is larger, representing the sum of the CPU time taken by each thread.
        Or a wall-clock time figure might be larger because it counts time spent waiting, while
        the corresponding system and user time figures only measure the time while the operation
        is actively using CPU cycles.
      </p>

      <p>
        The <xref href="impala_explain_plan.xml#perf_explain"><codeph>EXPLAIN</codeph> plan</xref> is also printed
        at the beginning of the query profile report, for convenience in examining both the logical and physical
        aspects of the query side-by-side. The
        <xref href="impala_explain_level.xml#explain_level">EXPLAIN_LEVEL</xref> query option also controls the
        verbosity of the <codeph>EXPLAIN</codeph> output printed by the <codeph>PROFILE</codeph> command.
      </p>
      <p>In <keyword keyref="impala32_full"/>, a new <codeph>Per Node
          Profiles</codeph> section was added to the profile output. The new
        section includes the following metrics that can be controlled by the
            <codeph><xref
            href="impala_resource_trace_ratio.xml#resource_trace_ratio"
            >RESOURCE_TRACE_RATIO</xref></codeph> query option.</p>
      <ul>
        <li><codeph>CpuIoWaitPercentage</codeph>
        </li>
        <li><codeph>CpuSysPercentage</codeph></li>
        <li><codeph>CpuUserPercentage</codeph></li>
        <li><codeph>HostDiskReadThroughput</codeph>: All data read by the host
          as part of the execution of this query (spilling), by the HDFS data
          node, and by other processes running on the same system.</li>
        <li><codeph>HostDiskWriteThroughput</codeph>: All data written by the
          host as part of the execution of this query (spilling), by the HDFS
          data node, and by other processes running on the same system.</li>
        <li><codeph>HostNetworkRx</codeph>: All data received by the host as
          part of the execution of this query, other queries, and other
          processes running on the same system. </li>
        <li><codeph>HostNetworkTx</codeph>: All data transmitted by the host as
          part of the execution of this query, other queries, and other
          processes running on the same system. </li>
      </ul>
      <!--AR 3/11/2019 The below example is out dated and does not add much value. Hiding it until this doc gets refactored.-->

      <p audience="hidden"> Here is an example of a query profile, from a
        relatively straightforward query on a single-node pseudo-distributed
        cluster to keep the output relatively brief. </p>

<codeblock audience="hidden">[localhost:21000] &gt; profile;
Query Runtime Profile:
Query (id=6540a03d4bee0691:4963d6269b210ebd):
  Summary:
    Session ID: ea4a197f1c7bf858:c74e66f72e3a33ba
    Session Type: BEESWAX
    Start Time: 2013-12-02 17:10:30.263067000
    End Time: 2013-12-02 17:10:50.932044000
    Query Type: QUERY
    Query State: FINISHED
    Query Status: OK
    Impala Version: impalad version 1.2.1 RELEASE (build edb5af1bcad63d410bc5d47cc203df3a880e9324)
    User: doc_demo
    Network Address: 127.0.0.1:49161
    Default Db: stats_testing
    Sql Statement: select t1.s, t2.s from t1 join t2 on (t1.id = t2.parent)
    Plan:
----------------
Estimated Per-Host Requirements: Memory=2.09GB VCores=2

PLAN FRAGMENT 0
  PARTITION: UNPARTITIONED

  4:EXCHANGE
     cardinality: unavailable
     per-host memory: unavailable
     tuple ids: 0 1

PLAN FRAGMENT 1
  PARTITION: RANDOM

  STREAM DATA SINK
    EXCHANGE ID: 4
    UNPARTITIONED

  2:HASH JOIN
  |  join op: INNER JOIN (BROADCAST)
  |  hash predicates:
  |    t1.id = t2.parent
  |  cardinality: unavailable
  |  per-host memory: 2.00GB
  |  tuple ids: 0 1
  |
  |----3:EXCHANGE
  |       cardinality: unavailable
  |       per-host memory: 0B
  |       tuple ids: 1
  |
  0:SCAN HDFS
     table=stats_testing.t1 #partitions=1/1 size=33B
     table stats: unavailable
     column stats: unavailable
     cardinality: unavailable
     per-host memory: 32.00MB
     tuple ids: 0

PLAN FRAGMENT 2
  PARTITION: RANDOM

  STREAM DATA SINK
    EXCHANGE ID: 3
    UNPARTITIONED

  1:SCAN HDFS
     table=stats_testing.t2 #partitions=1/1 size=960.00KB
     table stats: unavailable
     column stats: unavailable
     cardinality: unavailable
     per-host memory: 96.00MB
     tuple ids: 1
----------------
    Query Timeline: 20s670ms
       - Start execution: 2.559ms (2.559ms)
       - Planning finished: 23.587ms (21.27ms)
       - Rows available: 666.199ms (642.612ms)
       - First row fetched: 668.919ms (2.719ms)
       - Unregister query: 20s668ms (20s000ms)
  ImpalaServer:
     - ClientFetchWaitTimer: 19s637ms
     - RowMaterializationTimer: 167.121ms
  Execution Profile 6540a03d4bee0691:4963d6269b210ebd:(Active: 837.815ms, % non-child: 0.00%)
    Per Node Peak Memory Usage: impala-1.example.com:22000(7.42 MB)
     - FinalizationTimer: 0ns
    Coordinator Fragment:(Active: 195.198ms, % non-child: 0.00%)
      MemoryUsage(500.0ms): 16.00 KB, 7.42 MB, 7.33 MB, 7.10 MB, 6.94 MB, 6.71 MB, 6.56 MB, 6.40 MB, 6.17 MB, 6.02 MB, 5.79 MB, 5.63 MB, 5.48 MB, 5.25 MB, 5.09 MB, 4.86 MB, 4.71 MB, 4.47 MB, 4.32 MB, 4.09 MB, 3.93 MB, 3.78 MB, 3.55 MB, 3.39 MB, 3.16 MB, 3.01 MB, 2.78 MB, 2.62 MB, 2.39 MB, 2.24 MB, 2.08 MB, 1.85 MB, 1.70 MB, 1.54 MB, 1.31 MB, 1.16 MB, 948.00 KB, 790.00 KB, 553.00 KB, 395.00 KB, 237.00 KB
      ThreadUsage(500.0ms): 1
       - AverageThreadTokens: 1.00
       - PeakMemoryUsage: 7.42 MB
       - PrepareTime: 36.144us
       - RowsProduced: 98.30K (98304)
       - TotalCpuTime: 20s449ms
       - TotalNetworkWaitTime: 191.630ms
       - TotalStorageWaitTime: 0ns
      CodeGen:(Active: 150.679ms, % non-child: 77.19%)
         - CodegenTime: 0ns
         - CompileTime: 139.503ms
         - LoadTime: 10.7ms
         - ModuleFileSize: 95.27 KB
      EXCHANGE_NODE (id=4):(Active: 194.858ms, % non-child: 99.83%)
         - BytesReceived: 2.33 MB
         - ConvertRowBatchTime: 2.732ms
         - DataArrivalWaitTime: 191.118ms
         - DeserializeRowBatchTimer: 14.943ms
         - FirstBatchArrivalWaitTime: 191.117ms
         - PeakMemoryUsage: 7.41 MB
         - RowsReturned: 98.30K (98304)
         - RowsReturnedRate: 504.49 K/sec
         - SendersBlockedTimer: 0ns
         - SendersBlockedTotalTimer(*): 0ns
    Averaged Fragment 1:(Active: 442.360ms, % non-child: 0.00%)
      split sizes:  min: 33.00 B, max: 33.00 B, avg: 33.00 B, stddev: 0.00
      completion times: min:443.720ms  max:443.720ms  mean: 443.720ms  stddev:0ns
      execution rates: min:74.00 B/sec  max:74.00 B/sec  mean:74.00 B/sec  stddev:0.00 /sec
      num instances: 1
       - AverageThreadTokens: 1.00
       - PeakMemoryUsage: 6.06 MB
       - PrepareTime: 7.291ms
       - RowsProduced: 98.30K (98304)
       - TotalCpuTime: 784.259ms
       - TotalNetworkWaitTime: 388.818ms
       - TotalStorageWaitTime: 3.934ms
      CodeGen:(Active: 312.862ms, % non-child: 70.73%)
         - CodegenTime: 2.669ms
         - CompileTime: 302.467ms
         - LoadTime: 9.231ms
         - ModuleFileSize: 95.27 KB
      DataStreamSender (dst_id=4):(Active: 80.63ms, % non-child: 18.10%)
         - BytesSent: 2.33 MB
         - NetworkThroughput(*): 35.89 MB/sec
         - OverallThroughput: 29.06 MB/sec
         - PeakMemoryUsage: 5.33 KB
         - SerializeBatchTime: 26.487ms
         - ThriftTransmitTime(*): 64.814ms
         - UncompressedRowBatchSize: 6.66 MB
      HASH_JOIN_NODE (id=2):(Active: 362.25ms, % non-child: 3.92%)
         - BuildBuckets: 1.02K (1024)
         - BuildRows: 98.30K (98304)
         - BuildTime: 12.622ms
         - LoadFactor: 0.00
         - PeakMemoryUsage: 6.02 MB
         - ProbeRows: 3
         - ProbeTime: 3.579ms
         - RowsReturned: 98.30K (98304)
         - RowsReturnedRate: 271.54 K/sec
        EXCHANGE_NODE (id=3):(Active: 344.680ms, % non-child: 77.92%)
           - BytesReceived: 1.15 MB
           - ConvertRowBatchTime: 2.792ms
           - DataArrivalWaitTime: 339.936ms
           - DeserializeRowBatchTimer: 9.910ms
           - FirstBatchArrivalWaitTime: 199.474ms
           - PeakMemoryUsage: 156.00 KB
           - RowsReturned: 98.30K (98304)
           - RowsReturnedRate: 285.20 K/sec
           - SendersBlockedTimer: 0ns
           - SendersBlockedTotalTimer(*): 0ns
      HDFS_SCAN_NODE (id=0):(Active: 13.616us, % non-child: 0.00%)
         - AverageHdfsReadThreadConcurrency: 0.00
         - AverageScannerThreadConcurrency: 0.00
         - BytesRead: 33.00 B
         - BytesReadLocal: 33.00 B
         - BytesReadShortCircuit: 33.00 B
         - NumDisksAccessed: 1
         - NumScannerThreadsStarted: 1
         - PeakMemoryUsage: 46.00 KB
         - PerReadThreadRawHdfsThroughput: 287.52 KB/sec
         - RowsRead: 3
         - RowsReturned: 3
         - RowsReturnedRate: 220.33 K/sec
         - ScanRangesComplete: 1
         - ScannerThreadsInvoluntaryContextSwitches: 26
         - ScannerThreadsTotalWallClockTime: 55.199ms
           - DelimiterParseTime: 2.463us
           - MaterializeTupleTime(*): 1.226us
           - ScannerThreadsSysTime: 0ns
           - ScannerThreadsUserTime: 42.993ms
         - ScannerThreadsVoluntaryContextSwitches: 1
         - TotalRawHdfsReadTime(*): 112.86us
         - TotalReadThroughput: 0.00 /sec
    Averaged Fragment 2:(Active: 190.120ms, % non-child: 0.00%)
      split sizes:  min: 960.00 KB, max: 960.00 KB, avg: 960.00 KB, stddev: 0.00
      completion times: min:191.736ms  max:191.736ms  mean: 191.736ms  stddev:0ns
      execution rates: min:4.89 MB/sec  max:4.89 MB/sec  mean:4.89 MB/sec  stddev:0.00 /sec
      num instances: 1
       - AverageThreadTokens: 0.00
       - PeakMemoryUsage: 906.33 KB
       - PrepareTime: 3.67ms
       - RowsProduced: 98.30K (98304)
       - TotalCpuTime: 403.351ms
       - TotalNetworkWaitTime: 34.999ms
       - TotalStorageWaitTime: 108.675ms
      CodeGen:(Active: 162.57ms, % non-child: 85.24%)
         - CodegenTime: 3.133ms
         - CompileTime: 148.316ms
         - LoadTime: 12.317ms
         - ModuleFileSize: 95.27 KB
      DataStreamSender (dst_id=3):(Active: 70.620ms, % non-child: 37.14%)
         - BytesSent: 1.15 MB
         - NetworkThroughput(*): 23.30 MB/sec
         - OverallThroughput: 16.23 MB/sec
         - PeakMemoryUsage: 5.33 KB
         - SerializeBatchTime: 22.69ms
         - ThriftTransmitTime(*): 49.178ms
         - UncompressedRowBatchSize: 3.28 MB
      HDFS_SCAN_NODE (id=1):(Active: 118.839ms, % non-child: 62.51%)
         - AverageHdfsReadThreadConcurrency: 0.00
         - AverageScannerThreadConcurrency: 0.00
         - BytesRead: 960.00 KB
         - BytesReadLocal: 960.00 KB
         - BytesReadShortCircuit: 960.00 KB
         - NumDisksAccessed: 1
         - NumScannerThreadsStarted: 1
         - PeakMemoryUsage: 869.00 KB
         - PerReadThreadRawHdfsThroughput: 130.21 MB/sec
         - RowsRead: 98.30K (98304)
         - RowsReturned: 98.30K (98304)
         - RowsReturnedRate: 827.20 K/sec
         - ScanRangesComplete: 15
         - ScannerThreadsInvoluntaryContextSwitches: 34
         - ScannerThreadsTotalWallClockTime: 189.774ms
           - DelimiterParseTime: 15.703ms
           - MaterializeTupleTime(*): 3.419ms
           - ScannerThreadsSysTime: 1.999ms
           - ScannerThreadsUserTime: 44.993ms
         - ScannerThreadsVoluntaryContextSwitches: 118
         - TotalRawHdfsReadTime(*): 7.199ms
         - TotalReadThroughput: 0.00 /sec
    Fragment 1:
      Instance 6540a03d4bee0691:4963d6269b210ebf (host=impala-1.example.com:22000):(Active: 442.360ms, % non-child: 0.00%)
        Hdfs split stats (&lt;volume id&gt;:&lt;# splits&gt;/&lt;split lengths&gt;): 0:1/33.00 B
        MemoryUsage(500.0ms): 69.33 KB
        ThreadUsage(500.0ms): 1
         - AverageThreadTokens: 1.00
         - PeakMemoryUsage: 6.06 MB
         - PrepareTime: 7.291ms
         - RowsProduced: 98.30K (98304)
         - TotalCpuTime: 784.259ms
         - TotalNetworkWaitTime: 388.818ms
         - TotalStorageWaitTime: 3.934ms
        CodeGen:(Active: 312.862ms, % non-child: 70.73%)
           - CodegenTime: 2.669ms
           - CompileTime: 302.467ms
           - LoadTime: 9.231ms
           - ModuleFileSize: 95.27 KB
        DataStreamSender (dst_id=4):(Active: 80.63ms, % non-child: 18.10%)
           - BytesSent: 2.33 MB
           - NetworkThroughput(*): 35.89 MB/sec
           - OverallThroughput: 29.06 MB/sec
           - PeakMemoryUsage: 5.33 KB
           - SerializeBatchTime: 26.487ms
           - ThriftTransmitTime(*): 64.814ms
           - UncompressedRowBatchSize: 6.66 MB
        HASH_JOIN_NODE (id=2):(Active: 362.25ms, % non-child: 3.92%)
          ExecOption: Build Side Codegen Enabled, Probe Side Codegen Enabled, Hash Table Built Asynchronously
           - BuildBuckets: 1.02K (1024)
           - BuildRows: 98.30K (98304)
           - BuildTime: 12.622ms
           - LoadFactor: 0.00
           - PeakMemoryUsage: 6.02 MB
           - ProbeRows: 3
           - ProbeTime: 3.579ms
           - RowsReturned: 98.30K (98304)
           - RowsReturnedRate: 271.54 K/sec
          EXCHANGE_NODE (id=3):(Active: 344.680ms, % non-child: 77.92%)
             - BytesReceived: 1.15 MB
             - ConvertRowBatchTime: 2.792ms
             - DataArrivalWaitTime: 339.936ms
             - DeserializeRowBatchTimer: 9.910ms
             - FirstBatchArrivalWaitTime: 199.474ms
             - PeakMemoryUsage: 156.00 KB
             - RowsReturned: 98.30K (98304)
             - RowsReturnedRate: 285.20 K/sec
             - SendersBlockedTimer: 0ns
             - SendersBlockedTotalTimer(*): 0ns
        HDFS_SCAN_NODE (id=0):(Active: 13.616us, % non-child: 0.00%)
          Hdfs split stats (&lt;volume id&gt;:&lt;# splits&gt;/&lt;split lengths&gt;): 0:1/33.00 B
          Hdfs Read Thread Concurrency Bucket: 0:0% 1:0%
          File Formats: TEXT/NONE:1
          ExecOption: Codegen enabled: 1 out of 1
           - AverageHdfsReadThreadConcurrency: 0.00
           - AverageScannerThreadConcurrency: 0.00
           - BytesRead: 33.00 B
           - BytesReadLocal: 33.00 B
           - BytesReadShortCircuit: 33.00 B
           - NumDisksAccessed: 1
           - NumScannerThreadsStarted: 1
           - PeakMemoryUsage: 46.00 KB
           - PerReadThreadRawHdfsThroughput: 287.52 KB/sec
           - RowsRead: 3
           - RowsReturned: 3
           - RowsReturnedRate: 220.33 K/sec
           - ScanRangesComplete: 1
           - ScannerThreadsInvoluntaryContextSwitches: 26
           - ScannerThreadsTotalWallClockTime: 55.199ms
             - DelimiterParseTime: 2.463us
             - MaterializeTupleTime(*): 1.226us
             - ScannerThreadsSysTime: 0ns
             - ScannerThreadsUserTime: 42.993ms
           - ScannerThreadsVoluntaryContextSwitches: 1
           - TotalRawHdfsReadTime(*): 112.86us
           - TotalReadThroughput: 0.00 /sec
    Fragment 2:
      Instance 6540a03d4bee0691:4963d6269b210ec0 (host=impala-1.example.com:22000):(Active: 190.120ms, % non-child: 0.00%)
        Hdfs split stats (&lt;volume id&gt;:&lt;# splits&gt;/&lt;split lengths&gt;): 0:15/960.00 KB
         - AverageThreadTokens: 0.00
         - PeakMemoryUsage: 906.33 KB
         - PrepareTime: 3.67ms
         - RowsProduced: 98.30K (98304)
         - TotalCpuTime: 403.351ms
         - TotalNetworkWaitTime: 34.999ms
         - TotalStorageWaitTime: 108.675ms
        CodeGen:(Active: 162.57ms, % non-child: 85.24%)
           - CodegenTime: 3.133ms
           - CompileTime: 148.316ms
           - LoadTime: 12.317ms
           - ModuleFileSize: 95.27 KB
        DataStreamSender (dst_id=3):(Active: 70.620ms, % non-child: 37.14%)
           - BytesSent: 1.15 MB
           - NetworkThroughput(*): 23.30 MB/sec
           - OverallThroughput: 16.23 MB/sec
           - PeakMemoryUsage: 5.33 KB
           - SerializeBatchTime: 22.69ms
           - ThriftTransmitTime(*): 49.178ms
           - UncompressedRowBatchSize: 3.28 MB
        HDFS_SCAN_NODE (id=1):(Active: 118.839ms, % non-child: 62.51%)
          Hdfs split stats (&lt;volume id&gt;:&lt;# splits&gt;/&lt;split lengths&gt;): 0:15/960.00 KB
          Hdfs Read Thread Concurrency Bucket: 0:0% 1:0%
          File Formats: TEXT/NONE:15
          ExecOption: Codegen enabled: 15 out of 15
           - AverageHdfsReadThreadConcurrency: 0.00
           - AverageScannerThreadConcurrency: 0.00
           - BytesRead: 960.00 KB
           - BytesReadLocal: 960.00 KB
           - BytesReadShortCircuit: 960.00 KB
           - NumDisksAccessed: 1
           - NumScannerThreadsStarted: 1
           - PeakMemoryUsage: 869.00 KB
           - PerReadThreadRawHdfsThroughput: 130.21 MB/sec
           - RowsRead: 98.30K (98304)
           - RowsReturned: 98.30K (98304)
           - RowsReturnedRate: 827.20 K/sec
           - ScanRangesComplete: 15
           - ScannerThreadsInvoluntaryContextSwitches: 34
           - ScannerThreadsTotalWallClockTime: 189.774ms
             - DelimiterParseTime: 15.703ms
             - MaterializeTupleTime(*): 3.419ms
             - ScannerThreadsSysTime: 1.999ms
             - ScannerThreadsUserTime: 44.993ms
           - ScannerThreadsVoluntaryContextSwitches: 118
           - TotalRawHdfsReadTime(*): 7.199ms
           - TotalReadThroughput: 0.00 /sec</codeblock>
    </conbody>
  </concept>
</concept>