What influences the file size of Lightroom catalogue?

Question

Just a general question really...
I often read in various forums of LR users who have tens of thousands of images in their catalogue, and yet the catalogue file itself is relatively small (e.g. "I've got 100,000+ images in my catalogue, and it's 350MB). I personally have around 11,700 images in my catalogue at present, and the catalogue file is fast approaching the 8GB size (currently at around 7.69GB). To be clear, I'm talking specifically about the lrcat file here, not any associated previews. I have a resonable number of virtual copies (say around 100), and have added quite a bit of metadata (hundreds, if not thousands of keywords, GPS coordinates, captions, star ratings, flags and so on). I also carry out quite a few local adjustments to many - although by no means close to all - of my photos in the develop module and have some pretty large editing histories as a result. I also have a modest number of publishing services running, and multiple collections including a number of smart collections.
This is not a problem for me, and I have plenty of space for lrcat file to grow into so I'm certainly not losing sleep. I'm just intrigued as to why my catalogue file is (possibly) larger than it should be.
M

Answer 1

If you want to post (or send me) a link to the lrcat file, I'd take a look at it for you, and give you a break-down what's consuming all the bytes. But it might be fun to learn how to do that yourself (e.g. using SQL). I use SQLiteSpy, but other people have their favorites.. (or you can use a command-line client if you prefer..). One way: just run "drop table "{table-name}" on each table then look at filesize (do this to a copy, not the real thing).
Anyway, it's hard to imagine keywords and captions etc. taking much of the space, since even if you had 1000 10-character words of text metadata per photo average that still only adds up to 117MB, which isn't a substantial portion of that 8G you're seeing occupied.
Anyway, if you've painted the heck out of most of them and not cleared dev history, that'll do it - that's where I'd put my money too...
One thing to consider to keep file-size down:
===================================
* After reaching a milestone in your editing, take a snapshot then clear edit history, or the top part of it anyway (e.g. leave the import step), using a preset like:
Clear Edit History.lrtemplate
s = {
    id = "E36E8CB3-B52B-41AC-8FA9-1989FAFD5223",
    internalName = "No Edit",
    title = "Clear Edit History",
    type = "Develop",
    value = {
        settings = {
            NoEdit = true,
        uuid = "34402820-B470-4D5B-9369-0502F2176B7F",
    version = 0,
(that's my most frequently used preset, by far ;-})
PS - I've written a plugin called DevHistoryEditor, which can auto-consolidate steps and reduce catalog size - it's a bit cumbersome to use a.t.m. but in case you're interested...
Rob