- 最新
- 投票最多
- 评论最多
Yes, but I checked the data and there is no row with the format "...". It seems like Amazon does not acknowledge the rows after a certain amount and add and ellipsis (...) row on which it applies the query.
Have you really checked all records ? you can easily do it now with spark notebooks on athena.
I found out that some rows didn't fit the date_parse format by reviewing the files. I modified the request to exclude rows that don't fit the format with : SELECT cast(date_parse("date", '%d.%m.%Y %H:%i:%s') as timestamp) FROM "example_logs" WHERE "date" LIKE '0%' OR "date" LIKE '1%'OR "date" LIKE '2%'OR "date" LIKE '3%'. It means I only take rows where the first character is 0;1;2 or 3. It works because the specific lines where it doesn't fit, the data is corrupted and is of format Nan but I expect it is not the most robust solution.
The error you provided seems like there's a timestamp on your data that is not in the format you specified, so date_parse() will fail.