Hello. I’m a new user of DXO Photolab 9 and I am learning the philosophy of the tool.
I am trying to import a comma separated list of “n” keywords (output of an AI keywording online tool, a flat list, no hierarchies), to be associated to an image but I can’t. It seems like if it is possibile to add no more than a single keyword, repeating the process “n” times. This is not feasible, having to add about 50 keywords for each image, could you please suggest a solution ?
Welcome to the forum @mahatma63-mb
I’d try to
copy all keywords of the list and paste them into (the keyword field of) one image.
If the CSV separators are commas indeed, this should work.
If some other delimiter is used, replace it by a comma.
If the above does not fit your situation, describe it more clearly please. If you have one AI-generated CSV per image, the procedure will obviously be different. But again, I get too many degrees of freedom in interpreting your post.
Hello. I tried, but the string: keyword1,keyword2,…,keyword”n” is not split into “n” different keywords; it is imported as a single long string
Have you tied adding a space behind the comma?
not yet, when I’m back home I’ll try, thank you
Format should be a>b>c
And in the settings you have to configure if you want all the keywords a, b and c or just c with the hierarchy
Tried a few delimiters and they weren’t accepted.
Next test: make it an xmp sidecar. Maybe your AI online service can do that?
Reading the OP, I got the impression that it’s not about hierarchies, but about a flat list.
Haven’t used PL for keywording for quite some time, but it looks like I need to press ENTER after each keyword to make it work. I tried with returns between keywords, and PL accepts that…as a multi-line keyword.
@Joanna , any idea on how to convince PL to swallow text input?
Nope. This creates hierarchies and will right royally screw your database for future lookups.
Indeed. DxO have not thought this through at all, even after I provided them with a working solution during the beta of PL5. At present, the only delimiters that are accepted are | > <, which create hierarchies.
Here is a short screen movie of my app accepting a comma-delimited list of keywords…
To be honest, PhotoLab is not equipped for true metadata management, limited as it is to a tiny narrow column within the main window.
Here is my app, with its attached keyword manager…
As you can see, you can also import various keyword lists from delimited files and you can drag keywords from the browser into the keywords field for a multiple selection of images.
But, as @platypus mentions, this is only for Mac. But it is free on an “as is” basis with no guarantee. I have been using it for five years and allows me to safely work directly with RAW files as well as through XMP sidecars.
There is one way that will require some studying, which works on Windows as well. ExifTool is a command line tool that can do almost anything you want…
exiftool -sep ", " -keywords="One, Two, Three" *.JPG
This command will apply the comma separated words to all files with the JPG extension in the current folder.
ExifTool is an amazing and comprehensive tool and you would need to spend a couple of days getting to grips with it but there is a comprehensive help website available as well as a user forum where you can ask questions. It is also the engine behind my app.
Finally, choose a proper DAM app, which could be expensive and will do far more than you need.
Not exactly import, but note: hierarchy can type with ‘pipe’ separator ( | )
Example: A1|A2|A3|A4
p.s. i see @Joanna and others already write this.
May can you copy a few example line from your list? For curiosity, and its may add a few idea for me. May some SQL script can handle this.
Can you create file in structure like this example below (for each photo the related keyword(s) in separated row(s)) ?
| Filename | Keyword |
|---|---|
| Filename001.RAW | Keyword 01 |
| Filename001.RAW | Keyword 02 |
| Filename001.RAW | Keyword 03 |
| Filename002.RAW | Keyword 01 |
| Filename002.RAW | Keyword 22 |
| Filename002.RAW | Keyword 02 |
p.s.: i already write some SQL scripts for dump keyword list from PL9 - Manage keywords outside the database - #12 by andras.csore May write ‘import’ also can be done (i’m quite sure).
If we have indeed a load of couples like example.img and example.csv, the best bets might be exiftool and its learning curve or synthetic/edited sidecars, XMP or DOP, both of which aren‘t free of effort neither. Editing the database could be a way too, as mentioned by @andras.csore - and again - no free lunch.
Someone on speaking terms with AI could probably have it transform the csv into a compatible sidecar…
Flat keyword table (filtered) of a PL database on Mac:
Adding entries and linking these with the respective images needs some reverse engineering and SQL going beyond what I’m willing to try. It’s really sad that DxO doesn’t step in and finish all its half-baked asset and metadata management features. There’s more to those things than just adding rows in a DB.
Today is not possible to import a list of flat keyword list from any other tool since Photolab lacks an interface for that.
Lightroom for example is using TAB-separation and lists like that can be imported both to PhotoMechnaic and iMatch DAM in both directions..
The only thing really Photolab can do is to build a keyword list from the keywords already used on the pictures in Photolab or any other tool. That works fine.
DXO is aware of that feature request already but so far nothing has really happened.
The simplest way to use AI to add keywords to your pictures is to use iMatch DAM. Using ExifTools - good luck! If that is your choise it is better to do it manually in Photlab I guess.
I’m on the side of “don’t use PhotoLab for keywords”. It’s very basic on the management side. Technically, it can handle a flat keyword structure well, but I have found even if you do get the keywords in, the search isn’t intuitive.
I use Lightroom Classic to do all my keyword stuff. If I want to find photos I look in Lightroom. Then I know where they are stored and can go directly to them in PhotoLab. I suppose you might be able to use the integration between the two products to do that more automatically. I rarely need to work with more than a handful of “found” photos.
You can use Lightroom’s library module for free and, in my view, nothing beats it for features nor for ease of use.
Thank you all 
.
At last I decided to manage the keywords at DAM level, PL probably is not the right tool and every possible approach doesn’t seem to be fully satisfactory. The workflow will be a bit more complicated but I survive 
Hey! @platypus , @Stenis , @mahatma63-mb may you interested on that.
I write SQL script for keyword and keyword-photo relation import.
In Windows db version (i don’t have Mac)
Works fine (flat keywords).
Simple to use. It’s need to execute multiple steps, but still easy. Details commented in the Script.
KeywordImport_v1_0.zip (3.8 KB)
SQL script + sample CSV file
Hello everyone! If someone seriously interested, and need help / additional info on that, let me know.
@mahatma63-mb What AI stuff you use?
I’m using both OpenAI and Google Vision API integrated with DAM (I currently use Photo Supreme). The web tool I used, with online interface is: https://www.aistockkeywords.com/
@andras.csore, I ´m fine since I have solved that through using iMatch DAM the has all possibilities needed when it comes to handle imports/exports of keywordlists. and even automatically generate keywords through image analyzis.
But for some that just would like to say migrate from Lightroom to Photolab it can for sure be an alternative but if I should migrate from Lightroom I think I would turn on the export of IPTC/XMP and export all metadata from the Lightroom database and then just put all image folders under one single topfolder and just poiunt to that from Photolab and Indexes it inclusing all subfolders. Photolab will handle that with ease and that way auto-generate a matching keyword-list.
Tried it and found that RAW files aren’t accepted, at least not the ones from Canon.
Getting the keyword file for a JPEG took a few seconds only. The file also created a fairly adequate description plus the load of keywords mentioned in the OP. Too many for my taste though. Adding the contents of Encyclopedia Britannica to each image makes keywords useless 
The tool is not designed for raw files. My 80 MB Leica DNG would require a galactic broadband connection 
About the number of keywords they are just less than 50, the threshold typical of the Stock Agencies
