Resolving "Failed to parse input string" error in BigQuery

Question

I'm working on a cohort analysis where I count the number of users who signed up by week, then count the number of events each perform over subsequent weeks. (Pretty standard stuff!)

After spending a lot of time understanding BigQuery nesting and array data, I decided to create two views to flatten nested user_dim and event_dim columns using wildcard over the entire data set (tables generated of events on a daily basis). That's how I wound up with these views, on which my query is based:

USERS_VIEW schema

EVENT_VIEW schema

Query, and the Error Thrown

When I execute the following query joining those two views, I get the error "Failed to parse input string "20161111"":

    SELECT
    DATE_TRUNC(users.first_seen_date, WEEK) AS week,
    COUNT(DISTINCT users.uid) AS signed_up_users,
    COUNT(DISTINCT events.uid) AS logged_in_users,
    CASE
       WHEN COUNT(DISTINCT users.uid) > 0 THEN COUNT(DISTINCT events.uid) * 100 / COUNT(DISTINCT users.uid)
       ELSE 0
   END AS retention_pct
   FROM
      USERS_VIEW AS users
   LEFT JOIN
     EVENTS_VIEW AS events
    ON
      users.uid = events.uid
      AND PARSE_DATE('%x', events.event_date) >= DATE_ADD(users.first_seen_date, INTERVAL 1 WEEK)
      AND PARSE_DATE('%x', events.event_date) < DATE_ADD(users.first_seen_date, INTERVAL 2 WEEK)
     GROUP BY
      1
     ORDER BY
      1

I feel like this should be simple, but I can't figure out what formatting I'm missing to ensure the dates can be parsed. (And the UI doesn't tell me which line number is the offender.) I'm hoping it's a silly typo that someone else can see. Thanks in advance for any help!

Answer 1

I get the error "Failed to parse input string "20161111""

I think below will help to address that error

PARSE_DATE('%x', events.event_date)

should be

PARSE_DATE('%Y%m%d', events.event_date)

Also, optionally, you might want to change LEFT JOIN to just JOIN