Data should be a part of the version control for metatables
Some tables are metatables involving data/information connected to the software releases. The Source Control utility should have a possibility of setting a switch on the tables, turning data compare on. This way you could include data for these tables in the version control.
We plan on supporting comparison to data using Source Control this year.
I was wondering what format is best for the data as stored in source control, SQL, CSV, some oracle specific format.
What do people think?
Martine Vanasse commented
I nedd version control for table data with CLOB column; please support CLOB also . I think SQL is appropriate in that case but csv is fine too if it works.
Anrijs Vītoliņš commented
Another year passed. Any updates on having data compare with source control?
On format I would prefer SQL, but anything would work, which is easy to source-control.
XML would be worst option.
Pavel Voinov commented
Hi, Richard Mitchell.
"We plan on supporting comparison to data using Source Control this year"
2015 year is finished a long time ago. Do you still have a plan to release this feature?
Pavel Voinov commented
Any format which is easy compare in structural way is good.
XML and JSON are the best I think for such purpose
Will the planned work also include the restore of schema (default) data using SCO itself ??
That is the plan. I'm leaning towards using csv (or similar) storage on disk for metadata tables as parsing the table is much easier than the SQL approach.
The work is planned for this year.
Robert Evans commented
We need to be able to add certain tables to source control , then the redgate product to produce a script of changes i.e. insert, update and delete statements, based on the table primary key. The automated script would then be part of a release.
Mark Illum commented
my first instinct is sql insert statements. This would be similar to how objects are stored (sql create statements). It would also need to know/store which columns are ok to be different, like meaningless keys, and something like create date.
JC Park commented
Currently Data Compare for Oracle doesn't support comparison to scripts which would be a pre-requisite for this feature.
Iulian Ghinea commented
We need also this feature. A complete schema in SVN has ddls but also table data (constant data). This is necessary for a complete deploy.
We really need this one! Pretty please!
Olivier DALET commented
+3: Master Data tracking! We have the exact same need!
Sylvain Pointeau commented
Would be splendid, I struggle doing my own script to generate easily the master data. I hit a lot of border cases and I really missed a tool capable of it and bullet proof.