Edit: Discussion concluded and this is now a recommended improvement that is not really a new feature, but tech debt we should tackle
The current Lens table implementation is very simple and supports no interactivity or custom styling. There are three elements to the discussion of replacing the Lens table with a more powerful table:
- Are there product requirements that aren't supported by the basic table, but are supported by the data grid?
- Technical requirements for the table
- Consistency across Kibana
Unless there is a strong reason to use the data grid in one of these categories, Lens should not switch.
Product requirements that might need to be supported
Based on the table features for Lens:
| Feature |
Basic table |
Data grid |
| Changing sort order from table |
Lens is not using this feature, but could |
Yes |
| Align numbers to the right |
Lens is not using this feature, but could |
Yes |
| Reordering columns via drag & drop |
- |
Yes, but in a separate menu, WIP PR for in-header |
| Changing column width |
- |
Yes |
| Subtotal rows |
Yes, can add a footer row |
WIP PR |
| Hierarchical display |
- |
- |
| Grouped rows |
- |
- |
| Full pivot table support |
- |
- |
| Colored backgrounds |
Yes |
Yes |
| Infinite scroll |
Yes |
Not yet |
| Cell-level widgets |
Yes |
Yes |
The main difference is therefore support for changing column width, reordering columns, and subtotal rows.
Technical requirements
The EuiDataGrid must be performant and stable before we switch.
Consistency across Kibana
Currently, the EuiBasicTable is used in every table visualization that can be embedded in a dashboard, except for the upcoming Discover implementation. If we make a change only in Lens, will this introduce inconsistency?
Conclusion
Based on this analysis of the requirements for Lens tables, I do not think that the benefits outweigh the costs of switching.
cc @elastic/eui-design
Edit: Discussion concluded and this is now a recommended improvement that is not really a new feature, but tech debt we should tackle
The current Lens table implementation is very simple and supports no interactivity or custom styling. There are three elements to the discussion of replacing the Lens table with a more powerful table:
Unless there is a strong reason to use the data grid in one of these categories, Lens should not switch.
Product requirements that might need to be supported
Based on the table features for Lens:
The main difference is therefore support for changing column width, reordering columns, and subtotal rows.
Technical requirements
The EuiDataGrid must be performant and stable before we switch.
Consistency across Kibana
Currently, the EuiBasicTable is used in every table visualization that can be embedded in a dashboard, except for the upcoming Discover implementation. If we make a change only in Lens, will this introduce inconsistency?
Conclusion
Based on this analysis of the requirements for Lens tables, I do not think that the benefits outweigh the costs of switching.
cc @elastic/eui-design