SlickGrid -- viewing and grouping by multiple values field

Question

I'm using SlickGrid and very happy with it, one of the best grids I've ever encountered. I use DataView for the data and groupItemMetadataProvider for grouping.

I have a following problem and looking for a solution. My data contains some fields with multiple values. E.g. one data entry can be associated with multiple countries: US, IR, NZ or multiple categories: cats, dogs, birds.

Two questions:

• What is the best practice to implement the view for several entries in one cell. ? I suppose, custom template or formatter can be used here, but will be glad to hear your suggestions.
• How to implement grouping by such field? E.g. I would like to group by country, so the entry should be shown under each country group, same row under US, AU and NZ.

Real Life Example

Without grouping

Name    | Animals
-----------------
Paker   | cats, dogs, birds
Michael | cats, snakes

Group by animals

Name    | Animals
-----------------
- cats
Paker   | cats, dogs, birds
Michael | cats, snakes

- dogs
Paker   | cats, dogs, birds

- birds
Paker   | cats, dogs, birds

- snakes
Michael | cats, snakes

Cross-posted to SlickGrid Google Group: https://groups.google.com/forum/#!topic/slickgrid/BxS_4Lny3KE

Thanks a lot!

Answer 1

When displaying custom or collection based data within a single cell, if attempting to render that data vertically, you'll need to consider the lack of native support for dynamic row height.

Inspecting the grouping function (extractGroups ~line 472)

for (var i = 0, l = rows.length; i < l; i++) {
 r = rows[i];
 val = gi.getterIsAFn ? gi.getter(r) : r[gi.getter];
 group = groupsByVal[val];
 if (!group) {
   group = new Slick.Group();
   group.value = val;
   group.level = level;
   group.groupingKey = (parentGroup ? parentGroup.groupingKey + groupingDelimiter : '') + val;
   groups[groups.length] = group;
   groupsByVal[val] = group;
 }
 group.rows[group.count++] = r;
}

reveals that a custom getter is possible, but that a single value is used/expected in the grouping of each row. Thus, for a row having a multi-value cell to appear in multiple groups (without modyfying source), the data would have to be "flattened".

A less than ideal solution then becomes to maintain datasets respective to the view state being rendered. As data interaction needs and/or data size increase, this workaround becomes less feasible and tends toward the need to implement a customization in the source.