70-432:Managing Full Text indexes

While preparing for the 70-432 I’m digging in to Full Text indexing a bit more than I have before.  Today, I want to cover a fairly simple topic, when do you know you need to update your Full Text index?

Let’s cover a couple scenarios I’ve seen in the past that were solved by performing some maintenance on your Full Text indexes.

We just loaded a 2 GB file to the database

There are over 100 lbs of paper in the US Harmonized Trade ScheduleI used to work with international trade data  there is a lot of textual data in there.  Nearly all of it had to be searchable by our end users.  At least once a year every country would release updates to their rules and regulations.  There was no such thing as an update file, only drop all the data, then re-add it.

The problem with our Full Text indexes usually boiled down to the following report from our users:  “We searched for regulations on cotton imports, and we didn’t see any results.  I checked the source doc, and I can see them there.  Why aren’t they in the database, didn’t you load the new file?”

Whenever a new data is added to the database, and some of that data was loaded into a column with a Full Text index, you need to perform maintenance on that index.  Just like your regular indexes, they need the attention.  It’s just Full Text indexing feels different.  It isn’t.

Rebuild the index after such a large load, make sure you do a FULL sample, so that every row in the table is scanned for your Full Text index.

If you don’t want to have to deal with this manually every time, you could add it as the last step in your SSIS package that reads in that shiny new 2GB file.

We just finished making corrections for “country x” and now the search is acting weird.

Yeah, weird is a technical term.  When translated to IT speak it means It’s broke, it’s your fault, now fix it!

In the case where my end users told me they’d just finished making corrections for a country, it meant they’d read through all the imported data, and verified the imported version matched the printed version of the country’s regulations.  When they found a discrepancy, they’d update that record to make it jive with the print version of the regulations.

This usually ended up with about 10% of the records getting a tweak of some sort.

Again, you need to update the Full Text index.  But this time, run an INCREMENTAL population.  That way you don’t take the huge performance hit of updating the whole index, just those rows that were modified since the last population of your Full Text index.

Another note, Make this a nightly process, and save yourself the hassle of dealing with this manually.  If you have a set maintenance window, I’d suggest doing this just before the backup.  That way if you have to restore, you don’t have to follow the restore with an immediate Full Text operation.  It just saves time.

Well, I’m heading back to read more about Full Text indexing.  I know there is a lot more to pick up than I have.  If I see anything more that seems to jump out at me, I’ll share it later.  That way we’re both a little closer to getting the 70-432!

If you have any questions, send them in.  I’m here to help!

By Shannon Lowder

Shannon Lowder is the Database Engineer you've been looking for! Look no further for expertise in: Business Analysis to gather the business requirements for the database; Database Architecting to design the logical design of the database; Database Development to actually build the objects needed by the business logic; finally, Database Administration to keep the database running in top form, and making sure there is a disaster recovery plan.

Leave a comment

Your email address will not be published. Required fields are marked *