Question

我有一组带有一组值的表，表示例是

ID  |  Customer_name  | workorder
1   |    abc          | dispatch
2   |    xyz          | not_dispatch
3   |    jdk          | dispatch

并且这总共持续了1M行..现在我想将这个数据集采样到5000行，我希望3400个工作器为“not_dispatch”，1600个样本中带有“dispatch”。如何在PostgreSQL中完成。

Answer 1

远非有效但有效：

SELECT *
FROM (
  SELECT * FROM my_table
  WHERE workorder = 'dispatch' -- other filters
  ORDER BY random() LIMIT 1600) sub1
UNION
SELECT *
FROM (
  SELECT * FROM my_table
  WHERE workorder = 'not_dispatch' -- other filters
  ORDER BY random() LIMIT 3400) sub2;

对postgreSQL

1 个答案: