Simultaneous indexing?

[mulb-git]

Are there issues with indexing one location from 2 different indexing PCs simultaneously? Specifically for video files in this case.

[scrubbbbs]

There isn't anything to prevent this, but separate indexers cannot coordinate currently. It may work if you give each PC a different set of files or subdirectory to work on:

cbird -use <> -update <path> or
cbird -use <> -i.include <path> -i.include <path2> ... -update

For more info on filters read here

[mulb-git]

So I tried this and it mostly worked except I ignored the case in the directory paths (eg movies-p/divided/sci-fi instead of Movies-p/Divided/Sci-Fi and now i get the following message when I run an update:

@engine::update$ invalid removal, non-indexed path: "movies-p/divided/sci-fi/d9.mp4"

Is there a way to clear those out without re-indexing?

Other than some small issues, great utility :)

[scrubbbbs]

The include/exclude pattern is case-sensitive. What do you mean by "ignored case", is it something to do with how the filesystem is mounted? cbird does not modify the file name case.

Well you can always make corrections to the SQLite database file if you learn sql (or ask GPT)... sqlite.exe is included in the cbird directory. I can't tell exactly what is needed without some more information. But you would have something like:

sqlite.exe _index/media0.db # only file containing paths

select * from media where path like 'movies-p/%'; # should show nothing
select * from media where path like 'Movies-p/%'; # ??? does it show anything?
delete from media where path like '...'; # remove paths
update media set path='...' where path='...'; # modify paths

Note -v might tell you why cbird wants to remove it.

You can make a backup of _index/media.*.db and _index/videos in case something goes terribly wrong you won't have to re-index everything.

[mulb-git]

For context, I'm using Windows 11 - scanned drive is directly attached sata ntfs (4 disks in a raid 0). The folder names on the drive are capitalized and I ran the indexing command and used -i.include /movies-p/ instead of -i.include /Movies-p/

Now when I look at the results, it will show false positives where the paths will show that /movies-p/d9.mp4 and /Movies-p/d9.mp4 are duplicates.

[scrubbbbs]

used -i.include /movies-p/ instead of -i.include /Movies-p/

That's interesting. The -include option should not have matched /movies-p/ since everything is case-sensitive. So if you changed the capitalization of a folder it would be re-indexed...but that should have dropped everything with the old folder name. I'll have to take a look at this, seems like a bug. What values of -i.include did you use, only the lowercase or both?

If cbird -select-path "*movies-p*" -show is the same as cbird select-path "*Movies-p*" -show then these files were indexed twice... you can delete them from the index without dropping the other ones:

cbird-select path "*movies-p*" -remove

[mulb-git]

I ran i.include against a subdirectory and once all the scanning was done I ran cbird against the whole root folder which would have reindexed the files with the proper case, thus the two database entries for each file.

The select-path commands you listed show the same for both cases and the -remove flag worked to clean things up (although you've got a typo there) :)

Thanks again for the help.