如何在 BigQuery（标准 SQL）中实现“LIKE ANY”？答案

【问题标题】：How to implement "LIKE ANY" in BigQuery (standard SQL)?如何在 BigQuery（标准 SQL）中实现“LIKE ANY”？
【发布时间】：2019-07-05 20:08:24
【问题描述】：

我想使用“like any”功能。但是，BigQuery 无法识别“ANY”运算符。有什么功能可以代替“LIKE ANY”吗？

Bigquery - 标准 SQL

    SELECT
      date,
      CONCAT(fullvisitorid, CAST(visitid AS string)) ord_browsers,
      hits.appInfo.landingScreenName,
      channelgrouping,
      trafficSource.source,
      trafficSource.medium,
      CASE
        WHEN channelgrouping = 'seo_brand' THEN 'Branded'
        WHEN hits.appInfo.landingScreenName LIKE '%/article%'
      OR hits.appInfo.landingScreenName LIKE '%/p/%' THEN 'ADS'
        WHEN (hits.appInfo.landingScreenName LIKE ANY ('%/damesmode/%',  '%/herenkleding/%',  '%/kinderkleding/%',  '%/sportkleding/%')) THEN 'Plist'
        ELSE 'rest'
      END Landing_page
    FROM
      `meta-geography-145406.81187684.ga_sessions_*`,
      UNNEST(hits) AS hits
    WHERE
      (_table_suffix BETWEEN '20190201'
        AND '20190207')

语法错误：在 [12:47] 出现意外的关键字 ANY

【问题讨论】：

标签： google-bigquery

【解决方案1】：

同样问题的更简单查询：

WITH table AS (
  SELECT 'XXXX/herenkleding/XXX' x
)

SELECT x
FROM TABLE
WHERE x LIKE ANY ( 
  '%/damesmode/%'  '%/herenkleding/%',  '%/kinderkleding/%',  '%/sportkleding/%'
)

Syntax error: Unexpected keyword ANY at [2:6]

备选方案1：替换为一系列LIKE OR

SELECT x
FROM TABLE
WHERE (x LIKE '%/damesmode/%' 
OR x LIKE '%/herenkleding/%'
OR x LIKE '%/kinderkleding/%'
OR x LIKE'%/sportkleding/%')

XXXX/herenkleding/XXX

方案二：组合成一个正则表达式

WITH table AS (
  SELECT 'XXXX/herenkleding/XXX' x
)

SELECT x
FROM TABLE
WHERE REGEXP_CONTAINS(x
  , '/(damesmode|herenkleding|kinderkleding|sportkleding)/')

XXXX/herenkleding/XXX

【讨论】：

【解决方案2】：

这是另一种方法。这将保持您的原始查询完好无损，同时实现您想要的：

 SELECT
      date,
      CONCAT(fullvisitorid, CAST(visitid AS string)) ord_browsers,
      hits.appInfo.landingScreenName,
      channelgrouping,
      trafficSource.source,
      trafficSource.medium,
      CASE
        WHEN channelgrouping = 'seo_brand' THEN 'Branded'
        WHEN hits.appInfo.landingScreenName LIKE '%/article%'
      OR hits.appInfo.landingScreenName LIKE '%/p/%' THEN 'ADS'
        WHEN (hits.appInfo.landingScreenName LIKE c) THEN 'Plist'
        ELSE 'rest'
      END Landing_page
    FROM
      `meta-geography-145406.81187684.ga_sessions_*`,
      UNNEST(hits) AS hits, UNNEST([
       '%/damesmode/%',
       '%/herenkleding/%',
       '%/kinderkleding/%',
       '%/sportkleding/%']) AS c
    WHERE
      (_table_suffix BETWEEN '20190201'
        AND '20190207')

注意数组未嵌套，每个项目都在CASE 语句中进行比较。

【讨论】：

【解决方案3】：

以下是 BigQuery 标准 SQL

#standardSQL
SELECT
  DATE,
  CONCAT(fullvisitorid, CAST(visitid AS STRING)) ord_browsers,
  hits.appInfo.landingScreenName,
  channelgrouping,
  trafficSource.source,
  trafficSource.medium,
  CASE
    WHEN channelgrouping = 'seo_brand' THEN 'Branded'
    WHEN REGEXP_CONTAINS(hits.appInfo.landingScreenName , '/(article|p)/') THEN 'ADS'
    WHEN REGEXP_CONTAINS(hits.appInfo.landingScreenName , '/(damesmode|herenkleding|kinderkleding|sportkleding)/') THEN 'Plist'
    ELSE 'rest'
  END Landing_page
FROM `meta-geography-145406.81187684.ga_sessions_*`,
UNNEST(hits) AS hits
WHERE _TABLE_SUFFIX BETWEEN '20190201' AND '20190207'

【讨论】：