Question

我在下面的安排中给了我数据，该列下的时间戳似乎相同，但是无法以这样的方式实际导出数据，即在单个时间戳上仅给出一个时间戳柱。我该如何保持数据对的相关性？

    Timestamp, pizza shop, Timestamp, clothes store, Timestamp, stadium
    7/1/2019 0:00:00, 2, 7/1/2019 0:00:00, 4, 7/1/2019 0:00:00, 5
    7/1/2019 0:00:10, 3, 7/1/2019 0:00:10, 6, 7/1/2019 0:00:10, 7

Answer 1

您可以使用LookupAccountSid()将csv文件作为表挂载，然后使用它来选择另一个表。

例如：

CREATE EXTENSION file_fdw;

CREATE SERVER files FOREIGN DATA WRAPPER file_fdw;

CREATE FOREIGN TABLE csvfile (
  pizzashop_ts timestamp with time zone,
  pizzashop_count int,
  clothesstore_ts timestamp with time zone,
  clothesstore_count int,
  stadium_ts timestamp with time zone,
  stadium_count int)
SERVER files 
OPTIONS ( filename '/path/to/csv.file', format 'csv', header 'true');

然后我们从表中多次选择，但一次只选择列对，将它们与UNION ALL结合在一起，然后添加“ source”列以对数据进行分类：

SELECT pizzashop_ts AS "timestamp", pizzashop_count AS "count", 'pizza shop' AS source
FROM csvfile
UNION ALL
SELECT clothesstore_ts, clothesstore_count, 'clothes store' AS source
FROM csvfile
UNION ALL
SELECT stadium_ts, stadium_count, 'stadium' AS source
FROM csvfile;

这应该像这样布置您的CSV文件：

       timestamp        | count |    source
------------------------+-------+---------------
 2019-07-01 00:00:00+01 |     2 | pizza shop
 2019-07-01 00:00:10+01 |     3 | pizza shop
 2019-07-01 00:00:00+01 |     4 | clothes store
 2019-07-01 00:00:10+01 |     6 | clothes store
 2019-07-01 00:00:00+01 |     5 | stadium
 2019-07-01 00:00:10+01 |     7 | stadium
(6 rows)

自然地，您可以将其转换为INSERT查询，以将数据插入合适的表中。

披露：我是EnterpriseDB（EDB）员工。

从CSV将列数据对导入到SQL

1 个答案: