fix: alter sql columns to long text #7463
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
DiggidyDave
reviewed
May 7, 2019
| def upgrade(): | ||
| with op.batch_alter_table('query') as batch_op: | ||
| batch_op.alter_column( | ||
| 'sql', existing_type=sa.Text, type_=sa.Text(2 ** 32 - 1)) |
There was a problem hiding this comment.
Whoops, hooray for CI... postgresql is failing to migrate, looks like it has a 1GB limit on text. https://www.postgresql.org/docs/9.1/datatype-character.html
unclear to me if that GB is of the 2^30 or 10^9 variety... the latter/smaller one is probably enough for most people 🤣
mistercrunch
reviewed
May 7, 2019
| batch_op.alter_column( | ||
| 'executed_sql', existing_type=sa.Text, type_=sa.Text(length=text_len)) | ||
| except: | ||
| # Many databases do not have a length on text objects |
There was a problem hiding this comment.
I'd prefer having a MySQL-specific if block then a try silencing all errors
There was a problem hiding this comment.
Actually after reading https://docs.sqlalchemy.org/en/13/core/type_basics.html#sqlalchemy.types.Text I'm not sure either way. It's probably ok as is since catching a more specific exception might be hard because driver-specific...
There was a problem hiding this comment.
I would feel uncomfortable leaving a migration in an indeterminate state thought... this is a bit ugly but since there is divergent behavior between dbs for this column type, what about:
try:
do_upgrade(with_limit=True)
except:
do_upgrade(with_limit=False)
# if it fails after that, ok...
There was a problem hiding this comment.
Wait, are you suggesting that if it errors out then it didn't wan't/need the migration in the first place? That makes sense to me I think.
There was a problem hiding this comment.
Yes, if it errors out, it doesn't need the migration in the first place.
There was a problem hiding this comment.
Just a thought for the record: If we get rid of the batch_alter_table, it will fail after the first one and we will skip it gracefully. I think that's what we want. Not sure we need the batch here. A failure will fail as a batch anyway. Some really weird degenerate case would be if it is a mysql db and gets through 1 or 2 then the db crashes leaving one unmigrated. But that is not a catastrophic bad state.
There was a problem hiding this comment.
I ended up making the migration script only update the query table for mysql dbs. I know that it is very specific to Lyft's situation but I felt that my original implementation was not a good solution since every db has different requirements and data storage sizes for Text columns. So I don't think it is a good idea for me to try to set an arbitrary size and I wouldn't be surprised if it messed up for some other dbs.
Codecov Report
@@ Coverage Diff @@
## lyft-release-sp8 #7463 +/- ##
=================================================
Coverage 64.95% 64.95%
=================================================
Files 433 433
Lines 21357 21357
Branches 2352 2352
=================================================
Hits 13873 13873
Misses 7356 7356
Partials 128 128Continue to review full report at Codecov.
|
DiggidyDave
approved these changes
May 8, 2019
DiggidyDave
pushed a commit
to lyft/incubator-superset
that referenced
this pull request
May 8, 2019
Merge tip of lyft-release-sp8
DiggidyDave
pushed a commit
to lyft/incubator-superset
that referenced
this pull request
May 8, 2019
Merge lyft-develop to lyft-master
|
+1 |
khtruong
added a commit
to lyft/incubator-superset
that referenced
this pull request
May 8, 2019
Merge lyft-release-sp8@7bfe7bc to master
betodealmeida
pushed a commit
that referenced
this pull request
May 8, 2019
DiggidyDave
added a commit
to lyft/incubator-superset
that referenced
this pull request
May 8, 2019
fix: alter sql columns to long text apache#7463
DiggidyDave
added a commit
to lyft/incubator-superset
that referenced
this pull request
May 8, 2019
fix: alter sql columns to long text apache#7463
xtinec
pushed a commit
that referenced
this pull request
May 10, 2019
* filter out all nan series (#7313) * improve not rich tooltip (#7345) * Create issue_label_bot.yaml (#7341) * fix: do not save colors without a color scheme (#7347) * [wtforms] Strip leading/trailing whitespace (#7084) * [schema] Updating the datasources schema (#5451) * limit tables/views returned if schema is not provided (#7358) * limit tables/views returned if schema is not provided * fix typo * improve code performance * handle the case when table name or view name does not present a schema * Add type anno (#7342) * Updated local dev instructions to include missing step * First pass at type annotations * [schema] Updating the base column schema (#5452) * Update 937d04c16b64_update_datasources.py (#7361) * Feature flag for client cache (#7348) * Feature flag for client cache * Fix integration test * Revert "Fix integration test" This reverts commit 58434ab. * Feature flag for client cache * Fix integration tests * Add feature flag to config.py * Add another feature check * Fix more integration tests * Fix raw HTML in SliceAdder (#7338) * remove backendSync.json (#7331) * [bubbles] issue when using duplicated metrics (#7087) * SUPERSET-7: Docker compose config version breaks on Ubuntu 16.04 (#7359) * SUPERSET-8: Update text in docs copyright footer (#7360) * SUPERSET-7: Docker compose config version breaks on Ubuntu 16.04 * SUPERSET-8: Extra text in docs copyright footer * [schema] Adding commits and removing unnecessary foreign-key definitions (#7371) * Store last selected dashboard in sessionStorage (#7181) * Store last selected dashboard in sessionStorage * Fix tests * [schema] Updating the base metric schema (#5453) * Fix NoneType bug & fill the test recipients with original recipients if empty (#7365) * Added living goods as among the users of Superset (#7407) * Added living goods as among the users of Superset Living Goods is a non profit organisation with operation in africa and the middle east. We work in community health use data heavily on day to day. Superset is our platform of choice for dashboards. * Update README.md * [dashboard] allow user re-order top-level tabs (#7390) * [SQL Lab] Increase timeout threshold for offline check (#7411) * Bump FAB to 2.0.0 (#7323) * Bump FAB to 2.0.0 * [tests] whitelist SecurityApi login and refresh endpoints * [style] Fix, C812 missing trailing commas * [security] Remove SUPERSET_UPDATE_PERMS flag Registering sources needs to be performed after the views are initialized on UPDATE_PERMS=False configuration * [docs] New, FAB_UPDATE_PERMS and flask fab cli * [docs] Fix, db upgrade needs to come first, create-admin needs a db * [cli] New, superset init bootstraps all permissions for FAB and Superset * [style] Fix, flakes * [annotations] Improves UX on annotation validation, start_dttm, end_dttm (#7326) * Setting renderTrigger on label_colors (#7410) * Refactor out controlUtils.js module + unit tests (#7350) * [WiP]refactor out a controlUtils.js file * unit tests * add missing license * Addressing comments * feature: see Presto row and array data types (#7413) * Merge lastest from master into lyft-release-sp8 (#7405) * filter out all nan series (#7313) * improve not rich tooltip (#7345) * Create issue_label_bot.yaml (#7341) * fix: do not save colors without a color scheme (#7347) * [wtforms] Strip leading/trailing whitespace (#7084) * [schema] Updating the datasources schema (#5451) * limit tables/views returned if schema is not provided (#7358) * limit tables/views returned if schema is not provided * fix typo * improve code performance * handle the case when table name or view name does not present a schema * Add type anno (#7342) * Updated local dev instructions to include missing step * First pass at type annotations * [schema] Updating the base column schema (#5452) * Update 937d04c16b64_update_datasources.py (#7361) * Feature flag for client cache (#7348) * Feature flag for client cache * Fix integration test * Revert "Fix integration test" This reverts commit 58434ab. * Feature flag for client cache * Fix integration tests * Add feature flag to config.py * Add another feature check * Fix more integration tests * Fix raw HTML in SliceAdder (#7338) * remove backendSync.json (#7331) * [bubbles] issue when using duplicated metrics (#7087) * SUPERSET-7: Docker compose config version breaks on Ubuntu 16.04 (#7359) * SUPERSET-8: Update text in docs copyright footer (#7360) * SUPERSET-7: Docker compose config version breaks on Ubuntu 16.04 * SUPERSET-8: Extra text in docs copyright footer * [schema] Adding commits and removing unnecessary foreign-key definitions (#7371) * Store last selected dashboard in sessionStorage (#7181) * Store last selected dashboard in sessionStorage * Fix tests * [schema] Updating the base metric schema (#5453) * Fix NoneType bug & fill the test recipients with original recipients if empty (#7365) * feat: see Presto row and array data types (#7391) * feat: see Presto row and array data types * fix: address PR comments * fix: lint and build issues * fix: add types * add stronger type hints where possible * fix: lint issues and add select_star func in Hive * add missing pkg init * fix: build issues * fix: pylint issues * fix: use logging instead of print * Removed --console-log and superset runserver (#7421) * Fixes dashboard export button missing download and #7353 (#7427) * Added additional German translations to string file (#6604) * Added additional German translations to string file Updates to German translation files as per directions * Removed messages.json * [fix] Fixing SQL parsing issue (#7374) * add chinese translate (#7402) * Quick fix to address deadlock issue (#7434) * feat: view presto row objects in data grid (#7445) * Merge lastest from master into lyft-release-sp8 (#7405) * filter out all nan series (#7313) * improve not rich tooltip (#7345) * Create issue_label_bot.yaml (#7341) * fix: do not save colors without a color scheme (#7347) * [wtforms] Strip leading/trailing whitespace (#7084) * [schema] Updating the datasources schema (#5451) * limit tables/views returned if schema is not provided (#7358) * limit tables/views returned if schema is not provided * fix typo * improve code performance * handle the case when table name or view name does not present a schema * Add type anno (#7342) * Updated local dev instructions to include missing step * First pass at type annotations * [schema] Updating the base column schema (#5452) * Update 937d04c16b64_update_datasources.py (#7361) * Feature flag for client cache (#7348) * Feature flag for client cache * Fix integration test * Revert "Fix integration test" This reverts commit 58434ab. * Feature flag for client cache * Fix integration tests * Add feature flag to config.py * Add another feature check * Fix more integration tests * Fix raw HTML in SliceAdder (#7338) * remove backendSync.json (#7331) * [bubbles] issue when using duplicated metrics (#7087) * SUPERSET-7: Docker compose config version breaks on Ubuntu 16.04 (#7359) * SUPERSET-8: Update text in docs copyright footer (#7360) * SUPERSET-7: Docker compose config version breaks on Ubuntu 16.04 * SUPERSET-8: Extra text in docs copyright footer * [schema] Adding commits and removing unnecessary foreign-key definitions (#7371) * Store last selected dashboard in sessionStorage (#7181) * Store last selected dashboard in sessionStorage * Fix tests * [schema] Updating the base metric schema (#5453) * Fix NoneType bug & fill the test recipients with original recipients if empty (#7365) * feat: see Presto row and array data types (#7391) * feat: see Presto row and array data types * fix: address PR comments * fix: lint and build issues * fix: add types * Incorporate feedback from initial PR (prematurely merged to lyft-release-sp8) (#7415) * add stronger type hints where possible * fix: lint issues and add select_star func in Hive * add missing pkg init * fix: build issues * fix: pylint issues * fix: use logging instead of print * feat: view presto row objects in data grid * fix: address feedback * fix: spacing * feat: Scheduling queries from SQL Lab (#7416) (#7446) * Merge lastest from master into lyft-release-sp8 (#7405) * filter out all nan series (#7313) * improve not rich tooltip (#7345) * Create issue_label_bot.yaml (#7341) * fix: do not save colors without a color scheme (#7347) * [wtforms] Strip leading/trailing whitespace (#7084) * [schema] Updating the datasources schema (#5451) * limit tables/views returned if schema is not provided (#7358) * limit tables/views returned if schema is not provided * fix typo * improve code performance * handle the case when table name or view name does not present a schema * Add type anno (#7342) * Updated local dev instructions to include missing step * First pass at type annotations * [schema] Updating the base column schema (#5452) * Update 937d04c16b64_update_datasources.py (#7361) * Feature flag for client cache (#7348) * Feature flag for client cache * Fix integration test * Revert "Fix integration test" This reverts commit 58434ab. * Feature flag for client cache * Fix integration tests * Add feature flag to config.py * Add another feature check * Fix more integration tests * Fix raw HTML in SliceAdder (#7338) * remove backendSync.json (#7331) * [bubbles] issue when using duplicated metrics (#7087) * SUPERSET-7: Docker compose config version breaks on Ubuntu 16.04 (#7359) * SUPERSET-8: Update text in docs copyright footer (#7360) * SUPERSET-7: Docker compose config version breaks on Ubuntu 16.04 * SUPERSET-8: Extra text in docs copyright footer * [schema] Adding commits and removing unnecessary foreign-key definitions (#7371) * Store last selected dashboard in sessionStorage (#7181) * Store last selected dashboard in sessionStorage * Fix tests * [schema] Updating the base metric schema (#5453) * Fix NoneType bug & fill the test recipients with original recipients if empty (#7365) * feat: see Presto row and array data types (#7391) * feat: see Presto row and array data types * fix: address PR comments * fix: lint and build issues * fix: add types * Incorporate feedback from initial PR (prematurely merged to lyft-release-sp8) (#7415) * add stronger type hints where possible * fix: lint issues and add select_star func in Hive * add missing pkg init * fix: build issues * fix: pylint issues * fix: use logging instead of print * feat: view presto row objects in data grid * fix: address feedback * fix: spacing * Workaround for no results returned (#7442) * feat: view presto row objects in data grid (#7436) * feat: view presto row objects in data grid * fix: address feedback * fix: spacing * feat: Scheduling queries from SQL Lab (#7416) * Lightweight pipelines POC * Add docs * Minor fixes * Remove Lyft URL * Use enum * Minor fix * Fix unit tests * Mark props as required * feat: Add `validate_sql_json` endpoint for checking that a given sql query is valid for the chosen database (#7422) (#7462) merge from lyft-release-sp8 to master * Adds missing metric sum__SP_RUR_TOTL (#7452) * Late import for optional lib pyhive (#7471) * Late import for optional lib pyhive * fix * fix: calendar heatmap examples (#7375) Fixing a set of examples that trip on ValueError vs TypeError * bugfix: Improve support for special characters in schema and table names (#7297) * Bugfix to SQL Lab to support tables and schemas with characters that require quoting * Remove debugging prints * Add uri encoding to secondary tables call * Quote schema names for presto * Quote selected_schema on Snowflake, MySQL and Hive * Remove redundant parens * Add python unit tests * Add js unit test * Fix flake8 linting error * [dashboard] After update filter, trigger new queries when charts are visible (#7233) * trigger query when chart is visible * add integration test * fix: alter sql columns to long text #7463 (#7476) Merge lyft-release-sp8@7bfe7bc to master * Refactor ConsoleLog (#7428) * Revised Chinese translation (#7464) * add chinese translate * edit chinese translation * druid connector: avoid using 'dimensions' for scan queries (#7377) After the following PyDruid change (contained in version 0.5.2) the Superset Histogram charts rendered with Druid data are broken: druid-io/pydruid@0a59a70 Bump the pydruid requirements accordingly in setup.py Issue: #7368
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
CATEGORY
Choose one
SUMMARY
There were issues executing long queries because we saved the strings to text columns in the query table. The queries would be truncated based on the storage size of the text column (i.e. Text in mysql can hold 65k characters - 64 KB). I have added a migration script to update the columns to hold longer strings.
TEST PLAN
Tested that the columns were long text columns in a mysql table after migration.
REVIEWERS
@betodealmeida @xtinec @DiggidyDave