Start Building Professional
Web Apps Today

[dhtmlxGrid] Grouping Issues

Grouping seems to have some issues.

If you try the following steps on this online sample (http://dhtmlx.com/docs/products/dhtmlxGrid/samples/extended_modes/pro_group.html), grouping seems to loose his mind.

Reproducing steps :
- In a random row of the grid sample, rename author as 'j'.
- Choose another 3 random row and do the same.
- Choose a row not renamed and rename author as 'k'.
- Rename again author of the last row as 'j'.

Group 'j' show only now 3 rows and count 4. Javascript throw an error.
Seems to have same kind of issue with row edited with a dataprocessor on.

Thanks for you input on this.
N.D.

Problem confirmed and fixed. Updated js file sent by email. 

Issue was related to row moving logic, and was triggered by situation when last row in group need to be moved in group which is exactly above the row in question

File received and tested.
Issue is partly corrected as it introduces a new issue :

Reproducing steps :
-  Rename first row author as 'z'. Row goes at the bottom of the list.
- Rename same again as, in this example, 'Alexandr Pushkin'. Row goes in wrong group (just the one above him and not 'Alexandr Pushkin' group.

Furthermore, If  I delete rows (deleteRow(rowId) command), grouping breaks too in some cases (seems not related to current fix but already present in prior version).

I'm still investigating on this field for now, any input would be greatly appreciated.

Seems the issue is a bit more complicated , than it seems on first glance. 
We will made a more detailed testing of the problem and inform you when final fix will be available. 

It seems we was able to localize and fix all issues related to dyn. regrouping after edit operation. 
Fixed version of js file sent by email.

Seems to correct issue as far as my test goes.

However, after deleting rows (deleteRow(rowId) command), group  which was containing those rows 'breaks' (loose its 'group_row' div).
For now, workaround is to force grouping again after delete command.

Thanks for your dedicated attention anyway.