I am importing data from S3, one column is currently blank but I know that the column will contain dates (once populated by users).
However, I cannot set this column up now with a ‘yyyy-MM-dd’ format. Is there something I can do here? Looks like a bug.
Hi @charris - Can you check the data manually and see the sample data format. QuickSight is not able to identify the format the data type so a manual inspection will help or download the bad records and investigate the issues.
Regards - Sanjeeb
Hi, sorry I was not clear.
- The column is currently blank (it’s awaiting user input)
- It will contain dates in ‘yyyy-MM-dd’ format in the future
- In the dataset I cannot set the column up as a date in that format before it contains any data
I can only set it up as a string, which will be incorrect once data added.
QuickSight is expecting a date format to be entered as it considered it as date. you can put a format and during data set preparation change to string.
@Jesse - any advise on this.
Regards - Sanjeeb
Hi @charris - with S3 the schema is inferred from the content of the data, which is different than other data sources where the underlying source tells QuickSight what the column types are. Seems in this case without any values in the column you cannot convert it to a date yet. May need to wait for at least 1 row of data before you can convert. An alternative thing to try is to add a calculated field that uses parseDate() to result in a new date field. It should let you save the calc because the syntax will be valid.
There are many times rolling dashboards out to new clients that their data is not yet available. The parsedata could help, however the ability to define the data type directly would be far better.
This has been an issue before. I can mark this as a feature request.