Pegic is a user-facing command-line tool to read/write your data on Pegasus.
To connect to an existing Pegasus cluster, you need to specify the Pegasus address:
./pegic --meta 127.0.0.1:34601,127.0.0.2:34601
127.0.0.1:34601,127.0.0.2:34601 is a list of MetaServer addresses. It‘s a comma (’,') separated string.
Before starting read/write command, you need to firstly select a table to operate.
pegic » use xiaomi_ad_data
This command opens a table called “xiaomi_ad_data”. Every operation followed is applied to this table.
This is an overview of the commands that pegic provides.
Commands:
clear clear the screen
compression read the current compression algorithm
del delete a record
encoding read the current encoding
exit exit the shell
get read a record from Pegasus
help use 'help [command]' for command help
partition-index ADVANCED: calculate the partition index where the hashkey is routed to
scan scan records under the hashkey
set write a record into Pegasus
use select a table
pegic » set "hashkey::1" "sortkey" "value::1" ok Encoding: - HashKey: UTF8 - SortKey: UTF8 - Value: UTF8
Set/Get/Del correspond to “Write”,“Read” and “Delete” to a table. Each manipulates on only single record.
pegic » get "hashkey::1" "sortkey" hashkey::1 : sortkey : value::1 Encoding: - HashKey: UTF8 - SortKey: UTF8 - Value: UTF8
The first argument is the “hashkey”, which is used for partitioning. The second argument is the “sortkey”, that's key-ordering under the hashkey.
set command has the third argument, the “value”, which is the record body, the main data.
Pegasus stores the entire record in raw bytes, the database is unaware of the actual encoding of the data. By default, when you use pegic, the “hashkey”, “sortkey” and “value” will be encoded in UTF-8 from string to bytes.
pegic » set "中国" "北京" "abcdefg" ok Encoding: - HashKey: UTF8 - SortKey: UTF8 - Value: UTF8 pegic » get "中国" "北京" 中国 : 北京 : abcdefg Encoding: - HashKey: UTF8 - SortKey: UTF8 - Value: UTF8
However, the fields could also be set with other encodings like ByteBuffer.allocate(4).putInt(1695609641).array(), or be encoded via binary-serialization protocols like thrift/protobuf.
Here's an example where hashkey is encoded as an integer:
pegic » encoding hashkey int32 Encoding: - HashKey: INT32 - SortKey: UTF8 - Value: UTF8 pegic » get 19491001 "sortkey" 19491001 : sortkey : value Encoding: - HashKey: INT32 - SortKey: UTF8 - Value: UTF8
If the encoding is not correctly matched to the data, the result will be unexpected.
pegic » encoding hashkey utf8 Encoding: - HashKey: UTF8 - SortKey: UTF8 - Value: UTF8 pegic » get 19491001 "sortkey" Encoding: - HashKey: UTF8 - SortKey: UTF8 - Value: UTF8 error: record not found HASH_KEY=19491001 SORT_KEY=sortkey
utf8: The default encoding.
int32: 32-bit integer.
int64: 64-bit integer.
bytes: The golang byte array. Each byte is from range [0, 255]. For example: “104,101,108,108,111,32,119,111,114,108,100”.
javabytes: The java byte array. Each byte is from range [-128, 127]. For example: 49,48,52,48,56,10,0,3,-1,-1,0.
asciihex: For example \x00\x00\x00e\x0C\x00\x01\x08\x00\x01\x00\x00\x00e\x08.
Please submit an issue if there's encoding not supported but generally useful.
In some use-cases, we need to compress our value in order to improve read/write performance and latency stability.
For those data, you need to specify the compression algorithm in pegic:
pegic » compression set zstd ok pegic » get novelcoldstart novelcoldstart novelcoldstart : novelcoldstart : abcdef TTL(Time-To-Live) is not set Encoding: - HashKey: UTF8 - SortKey: UTF8 - Value: UTF8 Compression: zstd
pegic » scan "中国"
中国 : 上海 : {"people_num": 35000000}
中国 : 北京 : {"people_num": 40000000}
中国 : 广州 : {"people_num": 28000000}
中国 : 成都 : {"people_num": 20000000}
Total records count: 4
Encoding:
- HashKey: UTF8
- SortKey: UTF8
- Value: UTF8
scan iterates over all the records under the given hashkey. If you only want a subset of the records within a given range, or to filter out the records that matched some pattern, you can specify the flags according to your needs.
Scan with range:
pegic » scan --from 20200714000000 --to 20200715000000 uid18456112
uid18456112 : 20200714151429 : {"login": "Beijing"}
uid18456112 : 20200714152021 : {"login": "Beijing"}
uid18456112 : 20200714153342 : {"login": "Beijing"}
Total records count: 3
Encoding:
- HashKey: UTF8
- SortKey: INT64
- Value: UTF8
Scan with filtering pattern:
pegic » scan --prefix rocksdb pegasus pegasus : rocksdb_index_size_mb : 4300 pegasus : rocksdb_memtable_size_mb : 2000 pegasus : rocksdb_sst_size_mb : 358000 Total records count: 3 Encoding: - HashKey: UTF8 - SortKey: UTF8 - Value: INT64