4 ms·
I've done automated diff-based migrations before using SQLDelta. It caused me a lot of trouble and I think it's a bad approach. These diff tools just don't hav
by GlennS 6y ago
I've done automated diff-based migrations before using SQLDelta. It caused me a lot of trouble and I think it's a bad approach.
These diff tools just don't have the information they need to solve the data problems that are the actual hard part of schema migrations.
Simple example: I want to concatenate these two text columns into one.
I do think SQLDelta and similar tools can still be useful after doing the real migration using incremental scripts. You can use them to check for problems, and to fix up easy things like views and stored procedures.
- djrobstep 6y agoOf course they don't automatically solve data migrations for you, I'm not sure how any tool could possibly guess your intent like that. If you have an example of how a chain-based migration tool makes this easier, I'd love to hear about it.
- GlennS 6y agoI've never worked on anything that separated data migrations from schema migrations, so I suspect I'm not quite understanding you, but... Your data migration is just part of your chain of incremental patches? For my example you just write: ALTER TABLE my_table ADD COLUMN my_column TEXT NOT NULL DEFAULT ''; UPDATE TABLE my_table SET my_column = my_old_column_a || my_old_column_b; ALTER TABLE my_table ALTER COLUMN MY_COLUMN DROP DEFAULT;
- jashmatthews 6y agoBeyond a fairly small scale you have to separate data migrations from schema migrations because they just take too damn long.
- GlennS 6y agoI believe it and it makes sense, but I've never encountered it myself. Sounds like your definition of "small scale" encompasses every project I've ever worked on. :D
- djrobstep 6y agoRight, that's exactly what you might add to a diff-generated script to do the same thing. So it doesn't inherently seem better or worse to me.
- GlennS 6y agoThe difference is that - if I understood right - you're generating your diff just before you actually do the migration. By comparison, the incremental migration scripts are written at the same time as the feature (I like to do them as step 1 at lot of the time) and by the same developer, who currently has the problem in their head.
- djrobstep 6y agoI don't know of anybody doing that - unless you had automated checks that it was a fast and non-destructive operation, that would be highly risky. I'm simply talking about using the current production database as a basis for generating the required changes, rather than "here's what I think my production database currently looks like based on a version number and these 75 chained migration scripts I've run on it in the past"
- GlennS 6y agoI'm still confused. When do you generate your diff then? I think migrations ought to be done by the developer at the same time that they write their other code changes, because they have all the context at that time. However, it doesn't seem like that could work for diffs, because the version of the production database you're diffing against may not be the same as the version you end up deploying to.
- djrobstep 6y agoGenerate it whenever you prefer, probably with the rest of the code changes as you say. Immediately before applying, check that the database is still in the same state as it was when you generated the diff script. If yes, apply. If no, abort.