本文介绍了通过Batch Tool工具导入导出数据的方法。

工具介绍

Batch Tool工具是PolarDB-X团队开发的专门为PolarDB-X数据库提供数据导入导出服务的工具。

Batch Tool工具采用Java语言实现,基于生产者消费者模型并结合分布式数据库特点,实现了一站式且高效地从文件导入、导出到文件以及跨库的离线数据迁移(MySQL、PolarDB-X 1.0、PolarDB-X 2.0)等功能。在此基础上,还支持基于文本文件批量更新、删除等功能(实验性特性)。

工具获取

单击如下链接下载Batch Tool工具的jar包:

batch-tool.jar

参数介绍

命令行用法:java -jar batch-tool.jar --help

usage: BatchTool [-batchsize <size>] [-col <col1;col2;col3>] [-comp <NONE | GZIP>] [-con <consumer count>]
       [-config <filepath>] [-cs <charset>] [-D <database>] [-DDL <NONE | ONLY | WITH>] [-dir <directory
       path>] [-encrypt <NONE | AES | SM4>] [-error <max error count>] [-f <filepath1;filepath2>] [-F <file
       count>] [-fcon <parallelism>] [-format <NONE | TXT | CSV | XLS | XLSX>] [-func <true | false>] [-h
       <host>] [-H <filepath>] [-header <true | false>] [-help] [-i <true | false>] [-in <true | false>]
       [-initSqls <sqls>] [-key <string-type key>] [-L <line count>] [-lastSep <true | false>] [-lb <true |
       false>] [-local <true | false>] [-mask <Json format config>] [-maxConn <max connection>] [-maxWait
       <wait time(ms)>] [-minConn <min connection>] [-noEsc <true | false>] [-o <operation>] [-O <asc | desc>]
       [-OC <col1;col2;col3>] [-p <password>] [-P <port>] [-para <true | false>] [-param
       <key1=val1&key2=val2>] [-perf <true | false>] [-pre <prefix>] [-pro <producer count>] [-quote <AUTO |
       FORCE | NONE>] [-readsize <size(MB)>] [-rfonly <true | false>] [-ringsize <size (power of 2)>] [-s
       <separator char or string>] [-sharding <true | false>] [-t <tableName>] [-tps <tps limit>] [-u
       <username>] [-v] [-w <where condition>]
     -batchsize,--batchSize <size>                          Batch size of insert.
     -col,--columns <col1;col2;col3>                        Target columns for export.
     -comp,--compress <NONE | GZIP>                         Export or import compressed file (default NONE).
     -con,--consumer <consumer count>                       Configure number of consumer threads.
     -config,--configFile <filepath>                        Use yaml config file.
     -cs,--charset <charset>                                The charset of files.
     -D,--database <database>                               Database name.
     -DDL,--DDL <NONE | ONLY | WITH>                        Export or import with DDL sql mode (default NONE).
     -dir,--directory <directory path>                      Directory path including files to import.
     -encrypt,--encrypt <NONE | AES | SM4>                  Export or import with encrypted file (default NONE).
     -error,--maxError <max error count>                    Max error count threshold, program exits when the
                                                            limit is exceeded.
     -f,--file <filepath1;filepath2>                        Source file(s).
     -F,--filenum <file count>                              Fixed number of exported files.
     -fcon,--forceConsumer <parallelism>                    Configure if allow force consumer parallelism.
     -format,--fileFormat <NONE | TXT | CSV | XLS | XLSX>   File format (default NONE).
     -func,--sqlFunc <true | false>                         Use sql function to update (default false).
     -h,--host <host>                                       Host of database.
     -H,--historyFile <filepath>                            History file name.
     -header,--header <true | false>                        Whether the header line is column names (default
                                                            false).
     -help,--help                                           Help message.
     -i,--ignore <true | false>                             Flag of insert ignore and resume breakpoint (default
                                                            false).
     -in,--whereIn <true | false>                           Using where cols in (values).
     -initSqls,--initSqls <sqls>                            Connection init sqls (druid).
     -key,--secretKey <string-type key>                     Secret key used during encryption.
     -L,--line <line count>                                 Max line limit of one single export file.
     -lastSep,--withLastSep <true | false>                  Whether line ends with separator (default false).
     -lb,--loadbalance <true | false>                       Use jdbc load balance, filling the arg in $host like
                                                            'host1:port1,host2:port2' (default false).
     -local,--localMerge <true | false>                     Use local merge sort (default false).
     -mask,--mask <Json format config>                      Masking sensitive columns while exporting data.
     -maxConn,--maxConnection <max connection>              Max connection count (druid).
     -maxWait,--connMaxWait <wait time(ms)>                 Max wait time when getting a connection.
     -minConn,--minConnection <min connection>              Min connection count (druid).
     -noEsc,--noEscape <true | false>                       Do not escape value for sql (default false).
     -o,--operation <operation>                             Batch operation type: export / import / delete /
                                                            update.
     -O,--orderby <asc | desc>                              Order by type: asc / desc.
     -OC,--orderCol <col1;col2;col3>                        Ordered column names.
     -p,--password <password>                               Password of user.
     -P,--port <port>                                       Port number of database.
     -para,--paraMerge <true | false>                       Use parallel merge when doing order by export
                                                            (default false).
     -param,--connParam <key1=val1&key2=val2>               Jdbc connection params.
     -perf,--perfMode <true | false>                        Use performance mode at the sacrifice of compatibility
                                                            (default false).
     -pre,--prefix <prefix>                                 Export file name prefix.
     -pro,--producer <producer count>                       Configure number of producer threads (export /
                                                            import).
     -quote,--quoteMode <AUTO | FORCE | NONE>               The mode of how field values are enclosed by
                                                            double-quotes when exporting table (default AUTO).
     -readsize,--readSize <size(MB)>                        Read block size.
     -rfonly,--readFileOnly <true | false>                  Only read and process file, no sql execution (default
                                                            false).
     -ringsize,--ringSize <size (power of 2)>               Ring buffer size.
     -s,--sep <separator char or string>                    Separator between fields (delimiter).
     -sharding,--sharding <true | false>                    Whether enable sharding mode (default value depends on
                                                            operation).
     -t,--table <tableName>                                 Target table.
     -tps,--tpsLimit <tps limit>                            Configure of tps limit (default -1: no limit).
     -u,--user <username>                                   User for login.
     -v,--version                                           Show batch-tool version.
     -w,--where <where condition>                           Where condition: col1>99 AND col2<100 ...

命令主要包括以下类别:

  • 数据库连接配置
    • 基础连接信息:主机、端口、用户、密码等
    • 连接池配置:最大、最小连接数等
    • JDBC连接串参数
  • 批处理配置
    • 批处理功能参数
      • 待操作表名、文件名列表
      • 分隔符、是否以分隔符结尾、字符集、引号转义等文本文件读取或写入的相关参数
      • 文件数量、文件行数等导出配置
      • insert ingore、断点续传等导入配置
      • where、order by等sql条件
      • 压缩算法、加密算法、脱敏算法
      • 文件格式:csv、excel、txt等
    • 批处理性能参数
      • 生产者、消费者并行度设置
      • ringBuffer缓冲区、批数量、读取文件块等大小设置
      • pre partition、local merge等
      • tps限流相关

使用场景

在无特殊说明的情况下,下文中导入导出默认指定的字符集是utf-8,文件分隔符是,

  1. 假设需要导出tpch库下的customer表(分库分表模式)。
    1. 默认导出,文件数等于表的分片数:
      java -jar batch-tool.jar -P 3306 -h 127.0.XX.XX -u user_**** -p 12**** -D tpch -o export -t customer -s ,
    2. 导出为三个文件:
      java -jar batch-tool.jar -P 3306 -h 127.0.XX.XX -u user_**** -p 12**** -D tpch -o export -t customer -s , -F 3
    3. 导出为多个文件,单个文件最大行数为100000行:
      java -jar batch-tool.jar -P 3306 -h 127.0.XX.XX -u user_**** -p 12**** -D tpch -o export -t customer -s , -L 100000
    4. 指定where条件,默认导出:
      java -jar batch-tool.jar -P 3306 -h 127.0.XX.XX -u user_**** -p 12**** -D tpch -o export -t customer -s , -w "c_nationkey=10"
    5. 如果文本字段包含分隔符,则指定引号模式,默认导出:
      java -jar batch-tool.jar -P 3306 -h 127.0.XX.XX -u user_**** -p 12**** -D tpch -o export -t customer -s , -quote force
  2. 假设需要将csv文件导入到tpch库下的lineitem表(分库分表模式),其中对应库表已创建好。
    1. 指定单个文件导入:
      java -jar batch-tool.jar -P 3306 -h 127.0.XX.XX -u user_**** -p 12**** -D tpch -o import -t customer -s , -f "./data/lineitem.tbl"
    2. 指定文件夹路径下所有文件导入:
      java -jar batch-tool.jar -P 3306 -h 127.0.XX.XX -u user_**** -p 12**** -D tpch -o import -t customer -s , -dir "./data/lineitem/"