This package contains two tools for backing up PostgreSQL database dumps.
pg_dump_splitsort is a handy script for pre-processing PostgreSQL’s pg_dump output to make it more suitable for diffing and storing in version control.
The script splits the dump into the following files:
The files for table data are numbered so a simple sorted concatenation of all files can be used to re-create the database:
$ cat *.sql | psql <database>
I’ve found that a good way to take a quick look at differences between dumps is
to use the
meld tool on the whole directory:
$ meld old-dump/ new-dump/
Storing the dump in version control also gives a decent view on the differences. Here’s how to configure git to use color in diffs:
# ~/.gitconfig [color] diff = true [color "diff"] frag = white blue bold meta = white green bold commit = white red bold
Note: If you have created/dropped/renamed tables, remember to delete all
.sql files before post-processing the new dump.
The pg_incremental_backup script
pg_incremental_backup [-h] [--output-dir OUTPUT_DIR] database [remote] positional arguments: database remote optional arguments: -h, --help show this help message and exit --output-dir OUTPUT_DIR, -o OUTPUT_DIR