clickhouse 批量插入数据及ClickHouse常用命令详解

sudo yum install yum-utils
sudo rpm --import https://repo.yandex.ru/clickhouse/CLICKHOUSE-KEY.GPG
sudo yum-config-manager --add-repo https://repo.yandex.ru/clickhouse/rpm/stable/x86_64
sudo yum install clickhouse-server clickhouse-client
sudo /etc/init.d/clickhouse-server start
clickhouse-client

CLICKHOUSE_USER=username
 
CLICKHOUSE_LOGDIR=${CLICKHOUSE_HOME}/log/clickhoue-server
CLICKHOUSE_LOGDIR_USER=username
CLICKHOUSE_DATADIR_OLD=${CLICKHOUSE_HOME}/data/old
CLICKHOUSE_DATADIR=${CLICKHOUSE_HOME}/data

... ...
  <!-- 配置日志参数 -->
  <logger>
    <level>info</level>
    <log>${CLICKHOUSE_HOME}/log/clickhoue-server/clickhoue-server.log</log>
    <errorlog>${CLICKHOUSE_HOME}/log/clickhoue-server/clickhoue-server-error.log</errorlog>
    <size>100M</size>
    <count>5</count>
  </logger>
 
  <!-- 配置数据保存路径 -->
  <path>${CLICKHOUSE_HOME}</>
  <tmp_path>${CLICKHOUSE_HOME}/tmp</>
  <user_files_path>${CLICKHOUSE_HOME}/user_files</>
 
  <!-- 配置监听 -->
  <listen_host>::</listen_host>
 
  <!-- 配置时区 -->
  <timezone>Asiz/Shanghai</timezone>
... ...

#### a).启动服务
sudo service clickhouse-server start
#### b).停止服务
sudo service clickhouse-server stop

clickhouse-client

CREATE TABLE IF NOT EXISTS database.table_name ON cluster cluster_shardNum_replicasNum(
    'id' UInt64,
    'name' String,
    'time' UInt64,
    'age' UInt8,
    'flag' UInt8
)
ENGINE = MergeTree
PARTITION BY toDate(time/1000)
ORDER BY (id,name)
SETTINGS index_granularity = 8192

CREATE MATERIALIZED VIEW database.view_name ON cluster cluster_shardNum_replicasNum
ENGINE = AggregatingMergeTree
PARTITION BY toYYYYMMDD(time)
ORDER BY (id,name)
AS SELECT 
    toStartOfHour(toDateTime(time/1000)) as time,
    id,
    name,
    sumState( if (flag = 1, 1, 0)) AS successCount,
    sumState( if (flag = 0, 1, 0)) AS faildCount,
    sumState( if ((age < 10), 1, 0)) AS rang1Age,
    sumState( if ((age > 10) AND (age < 20), 2, 0)) AS rang2Age,
    sumState( if ((age > 20), 3, 0)) AS rang3Age,
    maxState(age) AS maxAge,
    minState(age) AS minAge
FROM datasource.table_name
GROUP BY time,id,name

INSERT INTO database.table_name(id, name, age, flag) VALUES(1, 'test', 15, 0)

INSERT INTO database.table_name FORMAT JSONEachRow{"id":"1", "name":"test", "age":"11", "flag":"1"}

SELECT * FROM database.table_name WHERE id=1

SELECT id, name, sumMerge(successCount), sumMerge(faildCount), sumMerge(rang1Age), sumMerge(rang2Age), maxMerge(maxAge), minMerge(minAge) 
FROM database.view_name 
WHERE id=1
GROUP BY id, name

CREATE TABLE IF NOT EXISTS database.table_name(
  'id' UInt64,
  'name' String,
  'time' UInt64,
  'age' UInt8,
  'flag' UInt8
nested_table_name Nested (
  sequence UInt32,
  id UInt64,
  name String,
  time UInt64,
  age UInt8,
  flag UInt8
  socketAddr String,
  socketRemotePort UInt32,
  socketLocalPort UInt32,
  eventTime UInt64,
  exceptionClassName String,
  hashCode Int32,
  nextSpanId UInt64
))
ENGINE = MergeTree
PARTITION BY toDate (time / 1000)
ORDER BY (id, name, time)
SETTINGS index_granularity = 8192

SELECT table1.*,table1.id FROM nest.table_name AS table1 array JOIN nested_table_name AS table2

<dictionaries>
  <dictionary>
    <name>url</name>  
    <source>
      <clickhouse>
        <host>hostname</host>  
        <port>9000</port>  
        <user>default</user>  
        <password/>  
        <db>dict</db>  
        <table>url_dict</table>
      </clickhouse>
    </source>  
    <lifetime>
      <min>30</min>  
      <max>36</max>
    </lifetime>  
    <layout>
      <hashed/>
    </layout>  
    <structure>
      <id>
        <name>id</name>
      </id>  
      <attribute>
        <name>hash_code</name>  
        <type>String</type>  
        <null_value/>
      </attribute>  
      <attribute>
        <name>url</name>  
        <type>String</type>  
        <null_value/>
      </attribute>
    </structure>
  </dictionary>  
  <dictionary>
    <name>url_hash</name>  
    <source>
      <clickhouse>
        <host>hostname</host>  
        <port>9000</port>  
        <user>default</user>  
        <password/>  
        <db>dict</db>  
        <table>url_hash</table>
      </clickhouse>
    </source>  
    <lifetime>
      <min>30</min>  
      <max>36</max>
    </lifetime>  
    <layout>
      <complex_key_hashed/>
    </layout>  
    <structure>
      <key>
        <attribute>
          <name>hash_code</name>  
          <type>String</type>
        </attribute>
      </key>  
      <attribute>
        <name>url</name>  
        <type>String</type>  
        <null_value/>
      </attribute>
    </structure>
  </dictionary>
</dictionaries>

SELECT
    id,
    dictGet('name', 'name', toUInt64(name)) AS name,
    dictGetString('url', 'url', tuple(url)) AS url
FROM table_name

clickhouse-client --query="INSERT INTO database.table_name FORMAT CSVWithNames" < /path/import_filename.csv

clickhouse-client --query="SELECT * FROM database.table_name FORMAT CSV" sed 's/"//g' > /path/export_filename.csv

SELECT table, name, partition，active FROM system.parts WHERE database='database_name'

ALTER TABLE database.table_name ON cluster cluster_shardNum_replicasNum detach partition 'partition_id'

SELECT
    database,
    table,
    name,
    formatReadableSize(sum(data_compressed_bytes) AS c) AS comp,
    formatReadableSize(sum(data_uncompressed_bytes) AS r) AS raw,
    c/r AS comp_ratio
FROM system.columns
WHERE database='database_name'
    AND table='table_name'
GROUP BY name

clickhouse-client --query="SELECT partition,count(*) AS partition_num, formatReadableSize(sum(bytes)) AS disk_size FROM system.columns WHERE database='database_name' " --external --?le=***.sql --name=parts --structure='table String, name String, partition UInt64, engine String' -h hostname