Version control systems can get hung up on large dump files, and opening large dump files to resolve conflicts can be virtually impossible.
One method could be to have an option to break up the database by tables and distribute the dumps amungst multiple files.
Comments
Comment #1
moshe weitzman commentedYes, I think a dump per table is a decent idea.
Comment #2
ceardach commentedNow dumping per table, and performing merges without any dump or restore. Overall, this should theoretically be tremendously faster on large databases. It will all depend upon the file system and speed of the system performing the merge. Theoretically, a merge on a database several gigabytes large should take roughly 20-40 minutes.