Question

我在将查询作为作业启动时遇到问题：

SELECT COUNT(*) FROM
(
  SELECT field_name FROM [dataset.table]
  WHERE 
    time BETWEEN DATE_ADD(CURRENT_TIMESTAMP(), -30, "DAY") AND
         CURRENT_TIMESTAMP() AND
  GROUP EACH BY field_name
) AS cur_month
JOIN EACH
(
   SELECT field_name FROM [dataset.table]
   WHERE 
     time BETWEEN DATE_ADD(CURRENT_TIMESTAMP(), -60, "DAY") AND
          DATE_ADD(CURRENT_TIMESTAMP(), -30, "DAY") AND
   GROUP EACH BY field_name
) AS prev_month
ON cur_month.field_name = prev_month.field_name

在控制台中运行此查询成功，但使用以下java代码运行它失败

JobConfigurationQuery queryConfig = new JobConfigurationQuery()
    .setQuery(query)
    .setDestinationTable(new TableReference()
            .setProjectId(projectId)
            .setDatasetId(toDataset)
            .setTableId(toTableId))
    .setAllowLargeResults(true)
    .setCreateDisposition("CREATE_IF_NEEDED")
    .setWriteDisposition("WRITE_TRUNCATE")
    .setPriority("BATCH")
    .setFlattenResults(false);

我得到的错误是

{
  "errorResult": {
    "location": "query",
    "message": "Ambiguous field name 'field_name' in JOIN. Please use the table qualifier before field name.",
    "reason": "invalidQuery"
  },
  "errors": [
    {
      "location": "query",
      "message": "Ambiguous field name 'field_name' in JOIN. Please use the table qualifier before field name.",
      "reason": "invalidQuery"
    }
  ],
  "state": "DONE"
}

有谁知道为什么？

Answer 1

问题与：BigQuery - same query works when submitted from UI and reports SQL syntax error from batch

重复

将展平结果设置为true可以解决我的问题。

Answer 2

您需要详细检查确实执行了您发布的查询，因为我在查询中看不到任何device_id，但错误消息很明确：

Ambiguous field name 'device_id' in JOIN. Please use the table qualifier before field name.

您需要确保device_id列在字段名称之前具有表限定符。

BigQuery可以在控制台中运行查询，但不能将其作为作业启动

2 个答案: