使用sql格式化字符串日期，例如“ Mon Sep 11 2017 00:00:00 GMT + 0200（CEST）”

Question

Well everything is in the title. How can I format a string with the following date format :

Mon Sep 11 2017 00:00:00 GMT+0200 (CEST)

It's in a Big Query Table within a standard SQL query.

I've started with this PARSE_DATE("%a %b %j %Y", Date) but I don't really know how to handle what's left...

thanks !

Answer 1

I'll start by simulating this type of data--everything uses CEST, but the days vary:

WITH `project.dataset.table` AS (
  SELECT
    FORMAT_DATE('%a %b %d %Y GMT+0200', date) AS start_date,
    FORMAT_DATE('%a %b %d %Y GMT+0200',
                DATE_ADD(date, INTERVAL MOD(off, 7) DAY)) AS end_date
  FROM UNNEST(GENERATE_DATE_ARRAY('2017-09-01', '2018-04-11', INTERVAL 7 DAY)) AS date WITH OFFSET off
)
SELECT
  start_date,
  end_date
FROM `project.dataset.table`;

This gives output that looks like Fri Oct 06 2017 00:00:00 GMT+0200 for start_date and Wed Oct 11 2017 00:00:00 GMT+0200 for end_date . Now to parse the date strings, we can use PARSE_DATE in conjunction with REGEXP_EXTRACT ; it's easiest to package the logic into a UDF for reuse in the query:

CREATE TEMP FUNCTION ConvertToDate(date_string STRING) AS (
  PARSE_DATE('%b %d %Y', REGEXP_EXTRACT(date_string, r'[^ ]+ (.*)00:00:00'))
);

Combining it with the sample input, we now have the following query:

CREATE TEMP FUNCTION ConvertToDate(date_string STRING) AS (
  PARSE_DATE('%b %d %Y', REGEXP_EXTRACT(date_string, r'[^ ]+ (.*)00:00:00'))
);

WITH `project.dataset.table` AS (
  SELECT
    FORMAT_DATE('%a %b %d %Y 00:00:00 GMT+0200', date) AS start_date,
    FORMAT_DATE('%a %b %d %Y 00:00:00 GMT+0200',
                DATE_ADD(date, INTERVAL MOD(off, 7) DAY)) AS end_date
  FROM UNNEST(GENERATE_DATE_ARRAY('2017-09-01', '2018-04-11', INTERVAL 7 DAY)) AS date WITH OFFSET off
)
SELECT
  ConvertToDate(start_date) AS start_date,
  ConvertToDate(end_date) AS end_date
FROM `project.dataset.table`;

This returns real DATE values for start_date and end_date , eg 2017-10-06 and 2017-10-11 . To apply this to your own table, remove the WITH clause and change the table name as appropriate.