[Severed Hand of Skywalker]

Just curious. What is it called and how can I find it to see how big it might be?

[real]

I was wondering that as well.
Anyone

[mactropolis]

Simple:
Check out the hidden folder .Spotlight-V100/ on the base of all your volumes. Inside that folder you'll find these files which constitute the index:

.journalHistoryLog
ContentIndex.db
store.db
.store.db
_rules.plist

PS. need to be root to access the .Spotlight-V100/ directory, not just an admin.

---
I Know It All

[CharlesS]

You can see how large your Spotlight index is by typing this:

sudo du -sh /.Spotlight-V100/

Mine is 369 MB.

[Millennium]

I would imagine that it depends largely upon how many files you have accessible to you. The size is unlikely to matter as much as the number, so a big drive with only a few files, such as one used for capturing video, would still have a smaller index than a small drive with many files, such as the drive you boot from.

[CatOne]

Originally Posted by Millennium

I would imagine that it depends largely upon how many files you have accessible to you. The size is unlikely to matter as much as the number, so a big drive with only a few files, such as one used for capturing video, would still have a smaller index than a small drive with many files, such as the drive you boot from.

This is going to be highly dependent on the file types. That's because Spotlight will do full text searches on "text" files (plain text, Word, PDF, every email, etc.) files, so if you have 10,000 1 MB text documents, you're going to have a HUGE file.

Larger files (videos, photos, etc.,) that aren't text, will just have the metadata indexed, so they won't add as much to the index.

[Millennium]

Originally Posted by CatOne

This is going to be highly dependent on the file types. That's because Spotlight will do full text searches on "text" files (plain text, Word, PDF, every email, etc.) files, so if you have 10,000 1 MB text documents, you're going to have a HUGE file.

The Project Gutenberg translation of Les Miserables, one of the longest single volumes around, is only three megs, and that file is pure ASCII. No one has that many 1-MB text files, if only because such files are few and far between outside of server logs. Also, it is unlikely that Spotlight actually stores the full text in its index. Instead, it ought to go out to the file itself, both to ensure that it has the most recent data and to save space. The slight performance overhead involved in opening the file will outweigh the potentially massive space (not to mention privacy and security) overhead of storing the file's contents in another location.

Most "text" files actually have very small amounts of actual text when compared to the size of the file. Word files are the worst when it comes to formatting bloat, but PDFs aren't exactly lean and mean either.

[Big Mac]

Originally Posted by Millennium

The Project Gutenberg translation of Les Miserables, one of the longest single volumes around, is only three megs, and that file is pure ASCII. No one has that many 1-MB text files, if only because such files are few and far between outside of server logs. Also, it is unlikely that Spotlight actually stores the full text in its index. Instead, it ought to go out to the file itself, both to ensure that it has the most recent data and to save space. The slight performance overhead involved in opening the file will outweigh the potentially massive space (not to mention privacy and security) overhead of storing the file's contents in another location.

Most "text" files actually have very small amounts of actual text when compared to the size of the file. Word files are the worst when it comes to formatting bloat, but PDFs aren't exactly lean and mean either.

Yeah, PDFs will definitely be larger than any native word processing file, including Word, since the format encapsulates a variety of information in order to display the file as it was seen in its native form. However, I don't see how that extra data would increase the size of the Spotlight index for that file type. As others have said, Spotlight is simply indexing text.

[stinch]

Originally Posted by Millennium

Also, it is unlikely that Spotlight actually stores the full text in its index. Instead, it ought to go out to the file itself, both to ensure that it has the most recent data and to save space. The slight performance overhead involved in opening the file will outweigh the potentially massive space (not to mention privacy and security) overhead of storing the file's contents in another location.

The performance overhead in opening the file and reading it to determine its contents isn't slight but massive.
Isn't the whole point to spotlight that almost all the info it needs about a file is stored in the database allowing fast searches?
Most text files people make are going to contain a small number of words repeated frequently. If the database can get away with storing a reference to a word and its frequency most documents will take up a lot less space in the database.

[Severed Hand of Skywalker]

I mostly have photoshop files. Anywho, mine is 148M.

[bmedina]

Spotlight will only index the first few hundred KB of files, IIRC.

[Big Mac]

Originally Posted by stinch

The performance overhead in opening the file and reading it to determine its contents isn't slight but massive.
Isn't the whole point to spotlight that almost all the info it needs about a file is stored in the database allowing fast searches?
Most text files people make are going to contain a small number of words repeated frequently. If the database can get away with storing a reference to a word and its frequency most documents will take up a lot less space in the database.

You're partially missing the point - it's an index rather than a database.

[CharlesS]

Originally Posted by Big Mac

You're partially missing the point - it's an index rather than a database.

Actually, the files within the invisible /.Spotlight-V100 directory have the extension .db. To me, that stands for "database."

[Person Man]

Originally Posted by CharlesS

Actually, the files within the invisible /.Spotlight-V100 directory have the extension .db. To me, that stands for "database."

It's an index stored in a database.

[CatOne]

Originally Posted by Millennium

The Project Gutenberg translation of Les Miserables, one of the longest single volumes around, is only three megs, and that file is pure ASCII. No one has that many 1-MB text files, if only because such files are few and far between outside of server logs. Also, it is unlikely that Spotlight actually stores the full text in its index. Instead, it ought to go out to the file itself, both to ensure that it has the most recent data and to save space. The slight performance overhead involved in opening the file will outweigh the potentially massive space (not to mention privacy and security) overhead of storing the file's contents in another location.

Most "text" files actually have very small amounts of actual text when compared to the size of the file. Word files are the worst when it comes to formatting bloat, but PDFs aren't exactly lean and mean either.

Milliennium, you're a software developer, right? Do you have Apple's developer documentation library on your machine? It's > 450 MB of PDFs. And much/most of that is text (sure, you can get a 120 page manual on the //vx if you want, and 5 copies of iMac technical specifications).

Also, I have > 100 MB of emails. And that's pretty much 100% text, because the attachments are in a separate folder. Many of those emails are > 80 kb in size, because people don't properly trim the emails on replies and they get long response chains. Gotta love sales reps.

Spotlight doesn't store the FULL text in its index (I'm sure it cuts out ifs, and, and buts). But it really does have to index many of the words in there... how else could you find "Pythagoras" in a PDF, without putting it in an index somewhere? Are you going to hash each word in some way?

[CatOne]

[turbo:/$] sudo du -sk .Spotlight-V100/
236716 .Spotlight-V100/