DRAFT: fix(dataframe): apply "Use time range" for non-INDEX TIMESTAMP columns #357
Conversation
| } | ||
|
|
||
| private constructTimeFilters(columns: Column[], timeRange: TimeRange): ColumnFilter[] { | ||
| const timeIndex = columns.find(c => c.dataType === 'TIMESTAMP' && c.columnType === 'INDEX'); |
There was a problem hiding this comment.
@atmgrifter00 @rajsite @m-akinc
Dashboard's global time range is used for decimation only if the TIMESTAMP is an INDEX column.
Note: Please read the PR description for more details.
I tried to find out why this condition was added. But as this was added as an initial commit by Carson Moore when the GIT Repo was created, I couldn't figure out the exact reason via PR history.
In Data Spaces, the filters are being applied even for the non-INDEX TIMESTAMP columns. Hence, I don't see any concerns in supporting the same here. But still, I would like to get your thoughts if I'm missing something.
I understand that there could be n number of columns with data type "TIMESTAMP" in a data table. Why didn't we add an option in the Data source to select the columns for which the time filters should get applied?
cc: @hemkumarrao-ni
Pull Request
🤨 Rationale
In the "SystemLink Data Frames" data source, the "Use time range" option works only when the TIMESTAMP column is an INDEX column.
But it should work even when the TIMESTAMP column is a non INDEX column.
Related bugs
Bug 3151208: Use time range does nothing when the timestamp column isn't the index column
Bug 3153240: Grafana data frame data source - Dashboard's global time range is used for decimation only if the time stamp is an index column
Systemlink Data Spaces
In Data spaces, even for non-INDEX TIMESTAMP columns, the time filters are applied.
👩💻 Implementation
Updates the functionality that constructs time filters to choose the TIMESTAMP column irrespective of the column type.
🧪 Testing
TBD
✅ Checklist