clickhouse输入输出格式 TSKV CSV

TSKV
TSKV格式不适合有大量小列的输出.
TSKV的效率并不比JSONEachRow差.
TSKV数据查询和数据导入。
不需要保证列的顺序。
支持忽略某些值，这些列使用默认值，例如0和空白行。复杂类型的值必须指定，无法使用默认值。

ch2 :) select * from escape_demo format TSKV;

SELECT *
FROM escape_demo
FORMAT TSKV

name=xiaomi    addr=nanjing\tjiangsu    age=23    desc=From nanjing
name=xiaohong    addr=zheng\tzhou    age=23    desc=A gril
name=xiaohong    addr=zheng\tzhou    age=23    desc=A gril

3 rows in set. Elapsed: 0.014 sec. 

[root@ch2 tmp]# clickhouse-client --query "select * from escape_demo format TSKV" > tskv.txt
[root@ch2 tmp]# cat tskv.txt
name=xiaomi    addr=nanjing\tjiangsu    age=23    desc=From nanjing
name=xiaohong    addr=zheng\tzhou    age=23    desc=A gril
name=xiaohong    addr=zheng\tzhou    age=23    desc=A gril

[root@ch2 tmp]# clickhouse-client --query "insert into escape_demo format TSKV" < tskv.txt
[root@ch2 tmp]# clickhouse-client
ClickHouse client version 20.3.9.70 (official build).
Connecting to localhost:9300 as user default.
Connected to ClickHouse server version 20.3.9 revision 54433.

ch2 :) select * from escape_demo;

SELECT *
FROM escape_demo

┌─name─────┬─addr───────────────┬─age─┬─desc─────────┐
│ xiaomi   │ nanjing    jiangsu │  23 │ From nanjing │
│ xiaohong │ zheng    zhou    │  23 │ A gril       │
│ xiaohong │ zheng    zhou    │  23 │ A gril       │
│ xiaomi   │ nanjing    jiangsu │  23 │ From nanjing │
│ xiaohong │ zheng    zhou    │  23 │ A gril       │
│ xiaohong │ zheng    zhou    │  23 │ A gril       │
└──────────┴────────────────────┴─────┴──────────────┘

CSV格式：CSV、CSVWithNames

1. CSV格式：CSV、CSVWithNames
CSV默认的分隔符为逗号，format_csv_delimiter设置自定义的分隔符。

CSV中的双引号使用两个双引号转义。

支持数据的查询和数据导入的。

create table csv_demo(create_date Date, update_time DateTime, desc String) ENGINE=TinyLog;

[root@ch2 tmp]# cat csv_demo.csv
2014-03-23|2014-03-23 14:10:14|Apache Spark achieves high performance
2014-03-23|2014-03-23 15:10:30|Spark offers over 80 high-level operators
1395990600|1395904200|Learning Apache "Spark" is easy

clickhouse-client --format_csv_delimiter="|" --query="insert into csv_demo format CSV" < csv_demo.csv

ch2 :) select * from csv_demo;

SELECT *
FROM csv_demo

┌─create_date─┬─────────update_time─┬─desc──────────────────────────────────────┐
│  2014-03-23 │ 2014-03-23 14:10:14 │ Apache Spark achieves high performance    │
│  2014-03-23 │ 2014-03-23 15:10:30 │ Spark offers over 80 high-level operators │
│  0000-00-00 │ 2014-03-27 07:10:00 │ Learning Apache "Spark" is easy           │
└─────────────┴─────────────────────┴───────────────────────────────────────────┘

3 rows in set. Elapsed: 0.004 sec. 

ch2 :) select * from csv_demo format CSV;

SELECT *
FROM csv_demo
FORMAT CSV

"2014-03-23","2014-03-23 14:10:14","Apache Spark achieves high performance"
"2014-03-23","2014-03-23 15:10:30","Spark offers over 80 high-level operators"
"0000-00-00","2014-03-27 07:10:00","Learning Apache ""Spark"" is easy"

3 rows in set. Elapsed: 0.004 sec. 

ch2 :) set format_csv_delimiter='|';

SET format_csv_delimiter = '|'

Ok.

0 rows in set. Elapsed: 0.002 sec. 

ch2 :) select * from csv_demo format CSV;

SELECT *
FROM csv_demo
FORMAT CSV

"2014-03-23"|"2014-03-23 14:10:14"|"Apache Spark achieves high performance"
"2014-03-23"|"2014-03-23 15:10:30"|"Spark offers over 80 high-level operators"
"0000-00-00"|"2014-03-27 07:10:00"|"Learning Apache ""Spark"" is easy"

3 rows in set. Elapsed: 0.004 sec. 

CSVWithNames可以自动过滤到CSV文件第一列，即名字列