This repository has been archived by the owner on Mar 24, 2021. It is now read-only.
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Format timestamp before creating id field
Slight differences in the way the _timestamp cell is formatted (in file uploads) results in duplicated data in the dataset. What’s happening is this: _timestamp is listed as an auto_id for the data set. All auto_id fields and values are joined together and then base64 encoded and stored in the _id field. AFTER the _id field has been created for the data record, the _timestamp is parsed and converted to utc format. This means that the date format of the timestamp can differ from the date format of the timestamp in the _id. The _id is used as a unique key for the data set. The same _timestamp with a different date format will generate a different _id value and will added as a new row in backdrop. The fix was just to move converting the _timestamp to utc format to BEFORE the base64 encoded _id is generated.
- Loading branch information