pg_bulkload使用记录

很久之前就使用过pg_bulkload来导入数据了，并做了对比试验，现在另一个项目又需要用了，这里做个记录：

1.rpm包比较老，下下来之后发现只支持到pg94，目前我用的是pg10，因此放弃。

2.下载源码安装：

git clone https://github.com/ossc-db/pg_bulkload.git

cd pg_bulkload

make && make install

--这里他会读取pg_config来获取pg的环境变量。

3.在要使用的数据库中执行：

create extension pg_bulkload;

4.导入csv文件：

pg_bulkload -i c_xxx.csv -O c_xxx -l c_xxx_load.log -d xxx -o "TYPE=CSV" -o "WRITER=PARALLEL"

5.导入压缩文件：

zcat c_xxx.gz |pg_bulkload -i stdin -O c_xxx -l c_xxx_load.log -d xxx -o "TYPE=CSV" -o "WRITER=PARALLEL"

6.关于-o的选项在help中没有，我们可以通过导入的log来看有哪些参数可以配置：

pg_bulkload 3.1.14 on 2018-09-28 11:31:12.641693+08

INPUT = stdin
PARSE_BADFILE = /var/lib/pgsql/pg10/data/pg_bulkload/20180928113112_sgdw_public_c_xxx.prs
LOGFILE = /var/lib/pgsql/sgdw/data/c_xxx_load.log
LIMIT = INFINITE
PARSE_ERRORS = 0
ENCODING = UTF8
CHECK_CONSTRAINTS = NO
TYPE = CSV
SKIP = 0
DELIMITER = ,
QUOTE = """
ESCAPE = """
NULL =
OUTPUT = public.c_xxx
MULTI_PROCESS = YES
VERBOSE = NO
WRITER = DIRECT
DUPLICATE_BADFILE = /var/lib/pgsql/pg10/data/pg_bulkload/20180928113112_sgdw_public_c_xxx.dup.csv
DUPLICATE_ERRORS = 0
ON_DUPLICATE_KEEP = NEW
TRUNCATE = YES


  0 Rows skipped.
  29423400 Rows successfully loaded.
  0 Rows not loaded due to parse errors.
  0 Rows not loaded due to duplicate errors.
  0 Rows replaced with new rows.

Run began on 2018-09-28 11:31:12.641693+08
Run ended on 2018-09-28 11:39:48.835205+08

CPU 2.63s/399.05u sec elapsed 516.19 sec

理论上黑体的都是可以配置的，比如配置为verbose为yes，那就在后面加一个-o "verbose=yes"

另外：默认逗号分隔，双引号将值括起来，默认直接写。如果忘记了，就导一个默认的，看看log就知道了。

附一个批量的脚本：

-bash-4.1$ cat load.sh
#!/bin/sh

#$1 data fil ename

file=$1

if [ ! -f $file  ]
then
    echo "File is not exist"
    exit 1
fi

echo "-----------------------------------------------------------------"

tbname=$( echo $file |cut -d . -f1 )
echo "Table name is : "$tbname

zcat $file|pg_bulkload -i stdin -O public.$tbname -l $tbname.log -o "TYPE=CSV" -o "WRITER=PARALLEL" -d sgdw

echo "load complete"
echo "-----------------------------------------------------------------"