Google BigQuery 时间戳比较

Question

there is a table "tbl1", having columns "name, surname, log_date". "log_date" is timestamp type.

How can I retrieve the records logged yesterday?

I am struggling with timestamp variable, I did something like:

declare yesterday timestamp
set (yesterday) = select TIMESTAMP_ADD(EXTRACT(Date FROM CURRENT_TIMESTAMP()), interval -1 day);

Above fails, so I am not able to use it later in my script:

SELECT distinct 
    name
FROM 
    `xxx.tbl1`
WHERE 
    log_date > yesterday

Also I tried:

SELECT 
    distinct name
FROM 
    `xxx.tbl1`
WHERE 
  log_date > TIMESTAMP_ADD(EXTRACT(Date FROM CURRENT_TIMESTAMP()), interval -1 day)

or

SELECT 
    distinct name
FROM 
    `xxx.tbl1`
WHERE 
  log_date > Select TIMESTAMP_ADD(EXTRACT(Date FROM CURRENT_TIMESTAMP()), interval -1 day)

or

WITH vars AS (
  SELECT EXTRACT(Select TIMESTAMP_ADD(EXTRACT(Date FROM CURRENT_TIMESTAMP()), interval -1 day) AS CustomDay
)
SELECT distinct 
    distinct name
FROM 
    `xxx.tbl1`, vars
WHERE 
    log_date > CustomDay

No luck.

Is there any other way, how to parse yesterday's date and use it in the select statement later on?

Answer 1

Without knowing the error you are getting, it seems that the problem is that you are trying to compare a timestamp (the log_date ) with datetime (the extracted DATE ). This should work:

SELECT 
    distinct name
FROM 
    `xxx.tbl1`
WHERE 
  DATE(log_date) = DATE_ADD(EXTRACT(DATE FROM CURRENT_TIMESTAMP()), INTERVAL -1 DAY)

Important: Use = for comparison to get the logs from yesterday only, > will get you the logs from today onwards.

One little thing: I would use DATE_SUB to subtract days instead of DATE_ADD with -1 (for readability), but kept it like that because of your example code.