Hi.

I want to rename my many photos to easily find all photos of a certain grandparent (not real names used in title). So any photos that only have 'nana', I want to find and rename manually to 'nana jane', etc.

I'd greatly appreciate a command line to achieve this.

Thanks.

The find utility can do boolean logic. See -a, -not, and -o in 'man find'. There is no XOR though. There is an implied AND between each option used when running find, so if you need an OR someday, you'll have to add in parenthesis to achieve the right grouping.

Then you could combine that with the rename utility, hopefully the perl-based one.

1 members found this post helpful.

find is too difficult when the temperature is above 35 degrees Celsius.

1 members found this post helpful.

or you can rename all nana to nana jane and afterward rename [back] nana jane joan to nana joan.
you only need to use the command rename to do that.

The existing answers do not give you what was actually asked for - they all wrongly exclude combinations like "joanna and nana" and wrongly include matches for results like "banana", and the obvious change would still wrongly exclude "nana joan and nana" from matching, which may or not occur but isn't difficult to cope with.

To search specifically for "nana" as a word and without jane or joan following it, you can use a regex with word boundary markers and negative lookahead: "\bnana\b(?! jane| joan)" or (useful if there might be more names): "\bnana\b(?! (jane|joan)\b)"

You can't use that with find because none of the 13 regex variants find allows have support for negative lookaheads, but you can easily filter with grep's Perl-compatible mode:

Or indeed using the recursive ls approach:
If you need to limit to JPEG files, the pattern should be suffixed with ".*\.jpe?g$" - but if these are scanned images they're as likely to be TIFFs or perhaps something else, so I've not bothered specifying the extension. If it's necessary you could use something like ".*\.(jpe?g|tiff?|tga|whatever)$"

Also, you probably want a case-insensitive search, so there's -i on the grep and -iname used for the find.

1 members found this post helpful.

Hi boughtonp.

That's the solution right there! Kudos and thank you!

Thanks to all for posting.

I don't have a solution, but it might be easier if you start with Ripgrep, which has full PCRE2 support.

Looking at that requirement I see that find is missing any options at all for recognizing word boundaries, not just negative lookaheads, in its regular expression pattern matching.

It is possible to use grep with find, though it might not be so efficient:

Hi TC.

Output:
I'll go with boughtonp's suggestion.

Not sure if the default regextype varies, but -regextype sed will allow \b word boundaries with find.

Some of the others engines don't support \b but do have \< and \> for start and end boundaries respectively.