将PostgreSQL数据库的表导入到elasticsearch中

Posted 2020-11-24 一泽涟漪

1.查看PostgreSQL表结构和数据信息

edbstore=# d customers
                                          Table "edbstore.customers"
        Column        |         Type          |                           Modifiers                            
----------------------+-----------------------+----------------------------------------------------------------
 customerid           | integer               | not null default nextval(‘customers_customerid_seq‘::regclass)
 firstname            | character varying(50) | not null
 lastname             | character varying(50) | not null
 address1             | character varying(50) | not null
 address2             | character varying(50) | 
 city                 | character varying(50) | not null
 state                | character varying(50) | 
 zip                  | integer               | 
 country              | character varying(50) | not null
 region               | smallint              | not null
 email                | character varying(50) | 
 phone                | character varying(50) | 
 creditcardtype       | integer               | not null
 creditcard           | character varying(50) | not null
 creditcardexpiration | character varying(50) | not null
 username             | character varying(50) | not null
 password             | character varying(50) | not null
 age                  | smallint              | 
 income               | integer               | 
 gender               | character varying(1)  | 
Indexes:
    "customers_pkey" PRIMARY KEY, btree (customerid)
    "ix_cust_username" UNIQUE, btree (username)
Referenced by:
    TABLE "cust_hist" CONSTRAINT "fk_cust_hist_customerid" FOREIGN KEY (customerid) REFERENCES customers(customerid) ON DELETE CASCADE
    TABLE "orders" CONSTRAINT "fk_customerid" FOREIGN KEY (customerid) REFERENCES customers(customerid) ON DELETE SET NULL

edbstore=# select count(1) from customers;
 count 
-------
 20000
(1 row)

2.利用PostgreSQL的row_to_json函数将表结构导出并保存为json格式

edbstore=# 	
Tuples only is on.
edbstore=# o customer.json
edbstore=# select row_to_json(r) from customers as r;
edbstore=# q

[postgres@sht-sgmhadoopcm-01 dba]$ ls -lh customer.json 
-rw-r--r-- 1 postgres appuser 7.7M Dec  7 22:37 customer.json

$ head -1 customer.json 
 {"customerid":1,"firstname":"VKUUXF","lastname":"ITHOMQJNYX","address1":"4608499546 Dell Way","address2":null,"city":"QSDPAGD","state":"SD","zip":24101,"country":"US","region":1,"email":"[e-mail]","phone":"4608499546","creditcardtype":1,"creditcard":"1979279217775911","creditcardexpiration":"2012/03","username":"user1","password":"password","age":55,"income":100000,"gender":"M"}

此时customer表虽然转储为json格式文件，但是并不能直接导入到elasticsearch，否则会报错如下

$ curl -H "Content-Type: application/json" -XPOST "172.16.101.55:9200/bank/_bulk?pretty&refresh" --data-binary "@customer.json"
{
  "error" : {
    "root_cause" : [
      {
        "type" : "illegal_argument_exception",
        "reason" : "Malformed action/metadata line [1], expected START_OBJECT or END_OBJECT but found [VALUE_NUMBER]"
      }
    ],
    "type" : "illegal_argument_exception",
    "reason" : "Malformed action/metadata line [1], expected START_OBJECT or END_OBJECT but found [VALUE_NUMBER]"
  },
  "status" : 400
}

 根据文档https://www.elastic.co/guide/en/elasticsearch/reference/current/docs-bulk.html说明，我们的json数据里并未指定每行数据唯一的文档id值

3.为json格式的表数据添加id字段

因为之前我们看到该customer表共有2000行，所以我们需要生成对应的20000个id值，我们借助python实现，新建build_id.py文件，并写入如下内容，看清楚是20001，因为包头不包尾原则，1-20000实际打印出来是1-19999，所以我们写1-20001

for i in range(1,20001):
    print(‘{"index":{"_id":"%s"}}‘ %i ) 

为该文件添加可执行权限，然后执行即可

$ python build_id.py > build_id.txt

$ head -3 build_id.txt 
{"index":{"_id":"1"}}
{"index":{"_id":"2"}}
{"index":{"_id":"3"}}

利用linux “paste"命令，将id文件和表文件合并

$ paste -d‘
‘ build_id.txt customer.json > customer_new.json

$ head -4 customer_new.json 
{"index":{"_id":"1"}}
 {"customerid":1,"firstname":"VKUUXF","lastname":"ITHOMQJNYX","address1":"4608499546 Dell Way","address2":null,"city":"QSDPAGD","state":"SD","zip":24101,"country":"US","region":1,"email":"[e-mail]","phone":"4608499546","creditcardtype":1,"creditcard":"1979279217775911","creditcardexpiration":"2012/03","username":"user1","password":"password","age":55,"income":100000,"gender":"M"}
{"index":{"_id":"2"}}
 {"customerid":2,"firstname":"HQNMZH","lastname":"UNUKXHJVXB","address1":"5119315633 Dell Way","address2":null,"city":"YNCERXJ","state":"AZ","zip":11802,"country":"US","region":1,"email":"[e-mail]","phone":"5119315633","creditcardtype":1,"creditcard":"3144519586581737","creditcardexpiration":"2012/11","username":"user2","password":"password","age":80,"income":40000,"gender":"M"}

 4.此时处理过的json格式的表文件就可以正常导入到elasticsearch中了，测试

$ curl -H "Content-Type: application/json" -XPOST "172.16.101.55:9200/customer/_bulk?pretty&refresh" --data-binary "@customer_new.json"

$ curl http://172.16.101.55:9200/_cat/indices?v
health status index    uuid                   pri rep docs.count docs.deleted store.size pri.store.size
yellow open   customer DvLoM7NjSYyjTwD5BSkK3A   1   1      20000            0       10mb           10mb