feat: add renameOperator #19776

zhaoyongjie · 2022-04-19T13:00:57Z

SUMMARY

Before Dataframe

Before PR, in the timeseries charts, the result column names follow the simple rules that combine all the levels into a single string. For example, in the birth_name dataset, select 1) ds as x-axis, 2) sum_num as metrics 3) gender as dimensions 4) 1990 : 2000 as time range 5) 1 year ago and actual values in time comparison. The intermediate MulitiIndex dataframe is:

             sum__num           sum__num__1 year ago          
gender            boy      girl                  boy      girl
ds                                                            
1990-01-01  1294971.0  899028.0            1281700.0  893001.0
1991-01-01  1248012.0  856255.0            1294971.0  899028.0
1992-01-01  1201680.0  820265.0            1248012.0  856255.0
1993-01-01  1154257.0  791723.0            1201680.0  820265.0
1994-01-01  1110122.0  760527.0            1154257.0  791723.0
1995-01-01  1072579.0  726740.0            1110122.0  760527.0
1996-01-01  1045266.0  702602.0            1072579.0  726740.0
1997-01-01  1021626.0  673936.0            1045266.0  702602.0
1998-01-01  1013188.0  657586.0            1021626.0  673936.0
1999-01-01   994617.0  636541.0            1013188.0  657586.0

then, the MultiIndex dataframe is changed plain dataframe by the flatten operator.

          ds  sum__num, boy  sum__num, girl  sum__num__1 year ago, boy  sum__num__1 year ago, girl
0 1990-01-01      1294971.0        899028.0                  1281700.0                    893001.0
1 1991-01-01      1248012.0        856255.0                  1294971.0                    899028.0
2 1992-01-01      1201680.0        820265.0                  1248012.0                    856255.0
3 1993-01-01      1154257.0        791723.0                  1201680.0                    820265.0
4 1994-01-01      1110122.0        760527.0                  1154257.0                    791723.0
5 1995-01-01      1072579.0        726740.0                  1110122.0                    760527.0
6 1996-01-01      1045266.0        702602.0                  1072579.0                    726740.0
7 1997-01-01      1021626.0        673936.0                  1045266.0                    702602.0
8 1998-01-01      1013188.0        657586.0                  1021626.0                    673936.0
9 1999-01-01       994617.0        636541.0                  1013188.0                    657586.0

The legacy NVD3 chart has a special rule that removes or renames columns. The rule will apply to columns if the chart has:

only has 1 metric
exists more than 1 series(breakdown in the line chart)
exists xAxis or TimeIndex
exists time comparison, and the comparison type is "actual values"

After the PR

So the PR introduces rename operator for this case

          ds        boy      girl  1 year ago, boy  1 year ago, girl
0 1990-01-01  1294971.0  899028.0        1281700.0          893001.0
1 1991-01-01  1248012.0  856255.0        1294971.0          899028.0
2 1992-01-01  1201680.0  820265.0        1248012.0          856255.0
3 1993-01-01  1154257.0  791723.0        1201680.0          820265.0
4 1994-01-01  1110122.0  760527.0        1154257.0          791723.0
5 1995-01-01  1072579.0  726740.0        1110122.0          760527.0
6 1996-01-01  1045266.0  702602.0        1072579.0          726740.0
7 1997-01-01  1021626.0  673936.0        1045266.0          702602.0
8 1998-01-01  1013188.0  657586.0        1021626.0          673936.0
9 1999-01-01   994617.0  636541.0        1013188.0          657586.0

Known limitations

Some derived metrics in time comparison are not addressed. the derived metrics refer to difference, percentage change, and radio in the time comparison.
There is no way customized column names after post processing.

I will handle these cases in the future and provide customized rename on the UI.

BEFORE/AFTER SCREENSHOTS OR ANIMATED GIF

After

TESTING INSTRUCTIONS

CI and before SUMMARY paragraph

ADDITIONAL INFORMATION

Has associated issue:
Required feature flags:
Changes UI
Includes DB Migration (follow approval process in SIP-59)
- Migration is atomic, supports rollback & is backwards-compatible
- Confirm DB migration upgrade and downgrade tested
- Runtime estimates and downtime expectations provided
Introduces new feature or API
Removes existing feature or API

codecov · 2022-04-19T13:29:03Z

Codecov Report

Merging #19776 (b14cef1) into master (231716c) will decrease coverage by 0.09%.
The diff coverage is 100.00%.

❗ Current head b14cef1 differs from pull request most recent head 2e4753e. Consider uploading reports for the commit 2e4753e to get more accurate results

@@            Coverage Diff             @@
##           master   #19776      +/-   ##
==========================================
- Coverage   66.51%   66.41%   -0.10%     
==========================================
  Files        1690     1691       +1     
  Lines       64730    64748      +18     
  Branches     6656     6660       +4     
==========================================
- Hits        43054    43003      -51     
- Misses      19977    20046      +69     
  Partials     1699     1699

Flag	Coverage Δ
hive	`?`
javascript	`51.23% <100.00%> (+0.02%)`	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
...superset-ui-core/src/query/types/PostProcessing.ts	`100.00% <ø> (ø)`
.../plugin-chart-echarts/src/Timeseries/buildQuery.ts	`66.66% <ø> (ø)`
...ui-chart-controls/src/operators/flattenOperator.ts	`100.00% <100.00%> (ø)`
...-ui-chart-controls/src/operators/renameOperator.ts	`100.00% <100.00%> (ø)`
...rontend/src/SqlLab/components/QueryTable/index.tsx	`70.17% <100.00%> (+0.53%)`	⬆️
superset/charts/schemas.py	`99.35% <100.00%> (ø)`
superset/utils/pandas_postprocessing/__init__.py	`100.00% <100.00%> (ø)`
superset/utils/pandas_postprocessing/flatten.py	`100.00% <100.00%> (ø)`
superset/views/core.py	`77.02% <100.00%> (+0.01%)`	⬆️
superset/db_engines/hive.py	`0.00% <0.00%> (-85.19%)`	⬇️
... and 7 more

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 231716c...2e4753e. Read the comment docs.

zhaoyongjie · 2022-04-20T07:59:14Z

superset/utils/pandas_postprocessing/flatten.py

+        for series in df.columns.to_flat_index():
+            _cells = []
+            for cell in series if is_sequence(series) else [series]:
+                if pd.notnull(cell):


The top-level columns are might set to None in the rename operator. Flatten skip the None level.

villebro

LGTM and tested to work well. Thanks so much for working on this! I think we're starting to approach the limits of what we can do with the current post processing functionality. To be able to move some of this functionality into the column/metric controls we will probably have to design a new API for these operations to make sure we're future proofed for any future use cases.

(cherry picked from commit 3c28cd4)

pull-request-size bot added the size/L label Apr 19, 2022

zhaoyongjie marked this pull request as draft April 19, 2022 13:15

zhaoyongjie added 3 commits April 20, 2022 10:57

wip

77f9c8c

rename operator

2fc9d1d

fix crash

a4cbd74

zhaoyongjie force-pushed the rename_operator branch from b486827 to a4cbd74 Compare April 20, 2022 02:57

zhaoyongjie added 3 commits April 20, 2022 12:22

remove unused operator

2c53f03

fix derived metrics

58c2da6

add ut

3db5304

pull-request-size bot added size/XL and removed size/L labels Apr 20, 2022

back port

651c523

zhaoyongjie marked this pull request as ready for review April 20, 2022 07:49

zhaoyongjie commented Apr 20, 2022

View reviewed changes

fix pylint

2e4753e

zhaoyongjie requested a review from a team April 20, 2022 09:18

villebro approved these changes Apr 20, 2022

View reviewed changes

zhaoyongjie merged commit 3c28cd4 into apache:master Apr 20, 2022

villebro added v1.5 lts-v1 labels Apr 20, 2022

zhaoyongjie mentioned this pull request Apr 21, 2022

Wrong data plotted when using Timeshift with GROUP BY columns populated in e-charts Time Series Line Chart #19741

Closed

villebro pushed a commit that referenced this pull request Apr 25, 2022

feat: add renameOperator (#19776)

8c8bbfb

(cherry picked from commit 3c28cd4)

hughhhh pushed a commit to hve-labs/superset that referenced this pull request May 11, 2022

feat: add renameOperator (apache#19776)

e336347

philipher29 pushed a commit to ValtechMobility/superset that referenced this pull request Jun 9, 2022

feat: add renameOperator (apache#19776)

d1ddc32

zhaoyongjie mentioned this pull request Jun 14, 2022

feat: adding truncate metric control on timeseries charts #20373

Merged

9 tasks

michael-s-molina removed the v1.5 label Jan 3, 2023

mistercrunch added 🍒 1.5.0 🍒 1.5.1 🍒 1.5.2 labels Mar 13, 2024

mistercrunch added 🍒 1.5.3 🏷️ bot A label used by `supersetbot` to keep track of which PR where auto-tagged with release labels 🚢 2.0.0 labels Mar 13, 2024

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: add renameOperator #19776

feat: add renameOperator #19776

zhaoyongjie commented Apr 19, 2022 •

edited

Loading

codecov bot commented Apr 19, 2022 •

edited

Loading

zhaoyongjie Apr 20, 2022

villebro left a comment

feat: add renameOperator #19776

feat: add renameOperator #19776

Conversation

zhaoyongjie commented Apr 19, 2022 • edited Loading

SUMMARY

Before Dataframe

After the PR

Known limitations

BEFORE/AFTER SCREENSHOTS OR ANIMATED GIF

After

TESTING INSTRUCTIONS

ADDITIONAL INFORMATION

codecov bot commented Apr 19, 2022 • edited Loading

Codecov Report

zhaoyongjie Apr 20, 2022

Choose a reason for hiding this comment

villebro left a comment

Choose a reason for hiding this comment

zhaoyongjie commented Apr 19, 2022 •

edited

Loading

codecov bot commented Apr 19, 2022 •

edited

Loading