Replace usage of in_array() in MigrateExecutable::handleMissingSourceRows #5766
Conversation
…urceRows(). Adds a new method getSourceIdKey() to serialize the source ID data into a unique key for retrieval later. Refer to drush-ops#5765
| */ | ||
| protected function getSourceIdKey(array $source_id): string | ||
| { | ||
| return implode('-', $source_id); |
There was a problem hiding this comment.
I wonder how safe is this
There was a problem hiding this comment.
I don't know the full extent of what is possible to appear in the $source_id array values, but from my investigation it is mostly entity IDs, revision IDs, and simple strings. Of which, most cases involve one or two values.
I am open to alternate ways of serializing the array values into a unique key.
There was a problem hiding this comment.
Hi.
I run into this issue while doing a diagnostic on what was slowing the import even after it was marked as finished.
And want to put my grain of salts.
The case i can see where the proposed serialisation can cause problem is very specific.
[A_B, C] has the same key than [A,B_C].
you could use hash function but you would still run into the Pigeon hole principle.
the only way to avoid that would be to let the user decide by a cli argument what separator to use, including the option to turn the feature off and use the old system.
PR to fix #5765