Support for Drupal 7 is ending on 5 January 2025—it’s time to migrate to Drupal 10! Learn about the many benefits of Drupal 10 and find migration tools in our resource center.
Problem/Motivation
The migration
process plugin never saves anything to the id map when a stub is created. Because of this, stubs are created that are never updated. :(
Proposed resolution
Add the source and destination to the id map so the stub can be updated at a later time.
Remaining tasks
Write Patch
User interface changes
N/A
API changes
N/A
Data model changes
Stubs created with the migration
process plugin will be added to the id map.
Comment | File | Size | Author |
---|---|---|---|
#30 | 2682705-30.patch | 2.78 KB | vasi |
#24 | 2682705-24.patch | 3.01 KB | alexpott |
#24 | 2682705.24.test-only.patch | 1.03 KB | alexpott |
#18 | migration-idmap-2682705-18.patch | 1.98 KB | vprocessor |
#3 | inner-diff.txt | 778 bytes | davidwbarratt |
Comments
Comment #2
davidwbarratt CreditAttribution: davidwbarratt at Golf Channel commentedComment #3
davidwbarratt CreditAttribution: davidwbarratt at Golf Channel commentedComment #4
davidwbarratt CreditAttribution: davidwbarratt at Golf Channel commentedComment #6
benjy CreditAttribution: benjy at PreviousNext commentedThis makes sense, I'm surprised it hasn't worked all this time, I did wonder if we should have an integration test but I guess the unit test is enough.
Comment #7
benjy CreditAttribution: benjy at PreviousNext commentedComment #8
catchI think this is worth an integration test as well - we should at least open a follow-up if not doing that here.
Comment #9
catchComment #10
benjy CreditAttribution: benjy at PreviousNext commentedI've closed #2639144: Stub row isn't saved in map as a duplicate because the patch there was red. But I asked the two contributors to comment here so they could be credited.
Comment #11
xjmRetitling to get at what (I think) is the implication of this.
Given how different Migrate is in 8.1.x vs. 8.0.x at this point let's make sure the patch works with 8.1.x.
Comment #12
xjmOkay I did not mean to queue three separate test runs but at least one of them is for 8.1.x. :)
Comment #14
xjmComment #15
vasi CreditAttribution: vasi commented@xjm, this is a problem for non-incremental migrations as well. It seriously affects any migrations related via eg: entity references. Eg, if nodes of type A point to taxonomy terms of type B, after migrating both I'll end up with:
* a bunch of garbage sample-data terms (created as stubs)
* nodes pointing to these garbage terms
* real terms, but nothing pointing to them
Comment #16
benjy CreditAttribution: benjy at PreviousNext commentedThis needs a re-roll against 8.1
Comment #17
vprocessor CreditAttribution: vprocessor at Skilld commentedComment #18
vprocessor CreditAttribution: vprocessor at Skilld commentedHi guys,
reroll is ready
Comment #19
vprocessor CreditAttribution: vprocessor at Skilld commentedComment #20
andypostComment #21
benjy CreditAttribution: benjy at PreviousNext commentedAs per #8, does anyone want to write an integration test?
Comment #22
catchTried an issue title update to reflect #15, which sounds like a good candidate for test coverage.
Comment #23
catchMarked #2690001: Migrate: Don't deduplicate forum vocabulary as duplicate.
Comment #24
alexpottHere's an addition to an integration test that proves it works... not in the migrate module though :(
Comment #26
benjy CreditAttribution: benjy at PreviousNext commentedLooks good.
Comment #29
catchCommitted/pushed to 8.2.x and 8.1.x, thanks!
This doesn't cherry-pick to 8.0.x, re-open if you port it.
miiimooo and milesw didn't comment here after 9 days, but added to the commit message.
Comment #30
vasi CreditAttribution: vasi commentedPorting to 8.0.x.
Comment #31
alexpottThis is in 8.1.x and 8.2.x therefore no longer a beta target.
Comment #32
xjmThanks @vasi.
I actually think we should not backport this, since 8.0.x is not a good development target for Migrate at the moment and is EOL in three weeks. So marking back to fixed for 8.1.x.