Advanced filtering


I have a new advanced search to do, only partly related to my previous one /t/16590/1

I need to filter all files to only those, which have the same beginning in the TITTLE tag, but only those cases where the is at least one with a little longer variation. And I dot know the names / titles

So for example I have files with titles [not filenames] like these:

Abc Def
Abc Def [Ghi]
Abc Def [Ghi Jkl]

Def Ghi
Ghi Jkl
GhiJkl [MNO]

What I need to find, in one and the same search [filtering], was made bold; what is to be disregarded [and not displayed] was made not. So, in a group of files like those above I only need to find the ones with the title "Abc Def" and its two longer variations- but not similar titles like "Abcdef" or "Def Ghi". And before search begins, I do not know if there is title like "Abc", "Abc Def", "Abc Def [Ghi]", "Def" etc.

If you want to have a comparison between different files and not different fields of a single file, then you will not find such a function.
MP3tag does not know anything about data in a file that is not the current file.
Yet, as there is the export function, you can generate a text file, load it e.g. into excel and let it compare the contents of the current field with that of the next field.

Thank you for the info

This situation sucks for me

I can use the Excel workaround but I have better: put the "[" in filter, copy the "Abc Def" from "Abc Def [Ghi]" and look if there is "Abc Def", given that I also have to listen to files with variations (and the "[" signs is crucial, beacause it means something in my naming system)

And even that workaround of mine will be time consuming, given that I have something like 3000 thousands of "[" files to check

In the Mp3tag list view, arrange the columns for filename and title, side by side.
Then sort one column.
Then compare the content of the columns within the rows, row by row.


You may try out an interactive heuristic approach by searching for similarities.
Script To Find Filenames Not Matching ID3 Tag
Simple Soundex Calculator


If I work on a new single album or even a whole artist folder- that's doable

Unfortunately, if I multiply that by thousands and hundreds, then my eyes are starting to pop out and eventually I'm missing some things, that I was suppose to find

Yes, quite. And along the way you will find loads of other files that do not comply with other standards.
But as any language is not logical, artists create names and titles that do not follow the rules and also the one who sets up a collection is not consistent in his use of wordings, any collection will evolve from random chaos to better structures. The first cut is the deepest.

You can create a Mp3tag export script, which does the main comparing work.
The export script should create a simple playlist text file, ...
having all the filepathnames, ...
where differences has been detected between the outer filename and the inner tag-fields.
Then load the new playlist into Mp3tag and do what to do next.

... or ...
Create an interactive Mp3tag Filter expression, which compares the filename with a formatstring of the involved tag-fields.


I have found a simple workaround, very good for my purposes

All I have to do is to:
1] put in filter "["
2] remove all tracks with empty ALBUM TAG [in that case less than 1% in my collection]
3] export names of the albums without repeating them
4] put TITLE tag in alphabetical order
5] open the exported list
6] copy name of one of those album into filter
7] look manually for tracks with "[Ghi]" variations [which are quite visible]
8] repeat step 6 and 7 until the exported list is done
9] take care manually of that omitted 1%

It's time consuming, because some albums will have only variations tracks and no "original" tracks without "[" in theirs titles, but in the end it works out the issues