Feature Request: Dreamcoder Style Grouping Interface

[peterlaursen]

If i understand you want a GUI-functionality for adding a GROUP BY clause to the last query?

If you have a table view in DATA TAB (what is populated by a SELECT * FROM TABLE) you want the GRID to populate with the result of  SELECT * FROM TABLE GROUP BY [SELECTED_COLUMN])

And also if you have a query result in REULT TAB (that could be populated by a SELECT col1, col2, col7 FROM TABLE WHERE IS NOT NULL) we should display the same as what SELECT col1, col2, col7 FROM TABLE WHERE IS NOT NULL GROUP BY[SELECTED_COLUMN]

And it could even be implemented iteratively so that we may  display what equals: SELECT * FROM TABLE GROUP BY [SELECTED_COLUMN1] DESC, [SELECTED_COLUMN2] ASC) .

Please confirm: do I understand correctly?

[peterlaursen]

BTW: do you know that you can sort data (“ORDER BY”) by clicking on the column header?

A GROUP BY with no aggregate function is rarely usable.  Best of all if you could tell what the SQL should be like.  If we have a TAB populated with “SELECT * FROM TABLE” what should the SQL look like when operating such functionality?

[NicholasA]

'peterlaursen' wrote on '06:

BTW: do you know that you can sort data (“ORDER BY”) by clicking on the column header?

A GROUP BY with no aggregate function is rarely usable.  Best of all if you could tell what the SQL should be like.  If we have a TAB populated with “SELECT * FROM TABLE” what should the SQL look like when operating such functionality?

You make a good point. I'm not sure if this is actually useful in helping you build queries, or whether it is just really useful in exploring and summarising your data in real-time. You are right, a true group-by would often have aggregate functions, whether this function actually just groups the actual records.

It is really useful from a data exploration point of view…a little like dragging and dropping fields in pivot tables (though again, this isn't a cross-tab, it's just grouping matching records).

So maybe this is just an interface and data exploration enhancement, rather than query building tool.

[Author]

Posts