PivotGrid - Column expands after datasource reload - T1246608 #28884
Conversation
Tucchhaa
force-pushed
the
T1246608_pivotgrid_header_resets_25_1
branch
from
ca2ca00 to
4f3d0bb
Compare
| loadOptions[expandedPathOptionName] = [[2010], [2010, 1]]; | ||
| store.load(loadOptions).done(function() { | ||
| assert.deepEqual(actualFilter, [['date', '=', 2010]], 'rows count'); | ||
| assert.deepEqual(actualFilter, ['date', '=', 2010], 'rows count'); | ||
| }); | ||
| }); |
There was a problem hiding this comment.
This test was to check if filter duplicates. Like [['date', '=', '2010'], 'or', ['date', '=', '2010']]. After my fix, the nesting structure is changed a little bit, but no duplicate filter occur, so that's ok
| store.load(loadOptions).done(function() { | ||
| assert.deepEqual(actualFilter, [[['date1', '=', 2010]], 'and', [['date2', '=', 2010]]], 'rows count'); | ||
| assert.deepEqual(actualFilter, [['date1', '=', 2010], 'and', ['date2', '=', 2010]], 'rows count'); | ||
| }); | ||
| }); |
There was a problem hiding this comment.
Same as this comment: https://github.com/DevExpress/DevExtreme/pull/28884/files#r1941724159
| @@ -1740,7 +1740,7 @@ QUnit.module('Expanding items', moduleConfig, () => { | |||
| { dataField: 'ShipRegion' }, | |||
| { dataField: 'ShipCity' } | |||
| ], | |||
| rowExpandedPaths: [['USA', 'AK']], | |||
| rowExpandedPaths: [['USA'], ['USA', 'AK']], | |||
There was a problem hiding this comment.
rowExpandedPaths: [['USA', 'AK']] is an impossible expanded path in pivotgrid. To open 'AK', you have to first open 'US', so the expandedPaths must be at least this: [['USA'], ['USA', 'AK']]
| rows, | ||
| filters: filters.concat(getExpandedPathSliceFilter(options, 'columns', j, firstCollapsedColumnIndex)).concat(rowFilterByExpandedPaths), | ||
| filters: commonFilters, |
There was a problem hiding this comment.
requestOptions.filters is an array of fields. getExpandedPathSliceFilter was extending fields with properties field.filterValues and 'field.filterType='include'. field.filterValues is an array of values of the field that we want to request. Later this array is transformed into a filter expression: https://github.com/DevExpress/DevExtreme/pull/28884/files#diff-b5d7b2b7d8a915d9d730cc006e8643b086ab02749e516b8cf9d46ec2dd1bd121L174
But now filters for fields are directly generated in the form of filter expression (aka [['field', '=', 'some value']]) and then just merged with options.filters: https://github.com/DevExpress/DevExtreme/pull/28884/files#diff-b5d7b2b7d8a915d9d730cc006e8643b086ab02749e516b8cf9d46ec2dd1bd121R205-R207
Tucchhaa
force-pushed
the
T1246608_pivotgrid_header_resets_25_1
branch
from
4f3d0bb to
d2f7078
Compare
Tucchhaa
force-pushed
the
T1246608_pivotgrid_header_resets_25_1
branch
from
d2f7078 to
b550e1e
Compare
| return 0; | ||
| } | ||
|
|
||
| function getExpandedPathsFilterExprByLevel(options, axis, level) { |
There was a problem hiding this comment.
this function fixes the issue. Before getExpandedPathSliceFilter was used to calculate filters.
Now the algorithm for calculating filters is different. getExpandedPathSliceFilter was extending fields array, adding filterValues and filterType=include to it. filterValues was an array which contained values of the columns that were expanded.
E.g. given this expandedPaths: [[2017], [2017, 1], [2018], [2018, 2]]. The result for level=2 will be this:
[
{ dataField: 'Year', filterValues: [2017, 2018], filterType: 'include' },
{ dataField: 'Quarter', filterValues: [1, 2], filterType: 'include' }
]
Later these fields will be converted into this filter expression:
[
[['Year', '=', '2017'], 'or', ['Year', '=', '2018]],
'and',
[['Quarter', '=', 1], 'or', ['Quarter', '=', 2]]
]
The problem with the resulted filter expression is that the data for the Quarter 2 of 2017 will be loaded (it matches the filter).
The new function, getExpandedPathsFilterExprByLevel, uses different algorithm and the resulting filter expression for the same expandedPaths for level=2 will be:
[
[['Year', '=', 1996], 'and', ['Month', '=', 1]],
'or',
[['Year', '=', 1997], 'and', ['Month', '=', 2]]
]
P.S. level is the level of nesting. PivotGrid has multi-level columns and rows. For each row level and column level a separate request is created.
No description provided.