Merge lp:~free.ekanayaka/storm/schema-sharding into lp:storm
Proposed by
Free Ekanayaka
Status: | Merged |
---|---|
Merged at revision: | 469 |
Proposed branch: | lp:~free.ekanayaka/storm/schema-sharding |
Merge into: | lp:storm |
Prerequisite: | lp:~free.ekanayaka/storm/schema-advance |
Diff against target: | 0 lines |
To merge this branch: | bzr merge lp:~free.ekanayaka/storm/schema-sharding |
Related bugs: |
Reviewer | Review Type | Date Requested | Status |
---|---|---|---|
Björn Tillenius (community) | Approve | ||
Review via email: mp+242133@code.launchpad.net |
Description of the change
Add a new storm.schema.
To post a comment you must log in.
Before reviewing the code itself, I'd like to have a discussion the approach it self.
Looking at the code, it looks like everytime I want to patch a single store, I would have to add an empty patch file for all the other stores? That seems cumbersome.
I'd like to have a single patch file that can patch all the stores. That will make it much easier to track what's going on when you have multiple-store patches. With different files it's hard to get an overview and you still have the problem that you have to know the order the stores get patched in.
For example, let's say that you have two stores, A and B. You want to patch both A and B, and migrate data using data from both A and B. If you select from tables you patched in A and B, you have to consider whether the table have been patched or not. It'd be much easier if you had everything in a single file, e.g.:
def patch(stores): something( stores[ "a"]) data(stores[ "a"], stores["b"]) something_ else(stores[ "b])
patch_
migrate_
patch_
What would be the downsides of having a single file for all the stores?