[bugfix] Fix Tableviz metrics column disorder #5566
Conversation
yamyamyuo
changed the title
Codecov Report
@@ Coverage Diff @@
## master #5566 +/- ##
=======================================
Coverage 63.82% 63.82%
=======================================
Files 364 364
Lines 23059 23059
Branches 2568 2568
=======================================
Hits 14717 14717
Misses 8327 8327
Partials 15 15
Continue to review full report at Codecov.
|
yamyamyuo
changed the title
|
Thanks for the fix! |
superset/viz.py
Outdated
| @@ -95,19 +95,21 @@ def __init__(self, datasource, form_data, force=False): | |||
| self.process_metrics() | |||
|
|
|||
| def process_metrics(self): | |||
| self.metric_dict = {} | |||
There was a problem hiding this comment.
Can't we just use an OrderedDict here? https://docs.python.org/2/library/collections.html#collections.OrderedDict
yamyamyuo
force-pushed
the
fix_tableviz_metrics_disorder
branch
from
5801ead
to
422b87a
Compare
yamyamyuo
force-pushed
the
fix_tableviz_metrics_disorder
branch
from
422b87a
to
d5e283b
Compare
yamyamyuo
force-pushed
the
fix_tableviz_metrics_disorder
branch
from
d5e283b
to
d365e01
Compare
|
@mistercrunch Any idea on a timeline for this PR? This is blocking our users and I hate having to build containers for things if I can just get 'em for free :) |
|
Merging this because of the unit test, but looks like the fix already made it into master. |
superset/viz.py
Outdated
| self.metric_dict[self.get_metric_label(o)] = o | ||
|
|
||
| # Cast to list needed to return serializable object in py3 | ||
| self.all_metrics = list(self.metric_dict.values()) |
There was a problem hiding this comment.
@mistercrunch this is the cause of metrics column disorder in TableViz.
It tries to use self.metric_dict and then converts to list here. I assume that it's going to remove duplicate metrics, since some metrics appear more than once in total which locate in different METRIC_KEYS .
superset/viz.py
Outdated
| # avoid generating duplicate metrics getting from METRIC_KEYS | ||
| if self.get_metric_label(o) not in self.metric_labels: | ||
| self.metric_labels.append(self.get_metric_label(o)) | ||
| self.all_metrics.append(o) |
There was a problem hiding this comment.
@mistercrunch so the best way to avoid duplicate metrics and keep its original order is using list at the beginning.
tests/viz_tests.py
Outdated
| u'optionName': u'metric_hww53ilkph_jo8b2fyt8rs'}, | ||
| u'SUM(SP_URB_TOTL)', | ||
| u'count'] | ||
| self.assertEqual(test_viz.all_metrics, expect_all_metrics) |
There was a problem hiding this comment.
now the labels and metrics order is correct.
superset/viz.py
Outdated
| self.metric_dict = {} | ||
| # metrics in TableViz is order sensitive, so metric_dict should be | ||
| # OrderedDict | ||
| self.metric_dict = OrderedDict() |
There was a problem hiding this comment.
@mistercrunch I have updated this self.metric_dict to OrderedDict() now. Please help to merge this PR.
* fix metrics disorder * add tableviz metric order tests * lint code * use OrderedDict to avoid metrics disorder * fix unit test
|
pivot_table can not order |
mistercrunch
added
🏷️ bot
Fixes #5373
The cause of metrics disorder in Table View is located in function
BaseViz. process_metrics.This function was introduced from https://github.com/apache/incubator-superset/pull/4914/files
This fix also works for percentage metrics.