Parsing RTF File

Hi:

I need to parse an regularly outputted rtf file and was wondering if it is possible in SSIS. I am trying to use the flat file connection manager to do this.

Now, I can't treat tab stops in an rtf like tab stops in a csv, since when you treat an rtf as a text file, you see the format code of the rtf. If I open the rtf in a text editor, the entire file is one line, with lines breaking with:

par}

Columns are tab delimited in the rtf, and they look like this when you treat the rtf as a text file.

plain abfs16f4cf0cb1

(or something like that, the word "tab" is the important part.)

So I use the "plain ab" part to delimit in SSIS, since that is consistent (planning to parse out all the garbage later on). The problem is, sometimes lines don't have a "city" and "state", so it "tabs" right over to the next field. So like this (looking in MS Word):

Phone <tab> City <tab> State <tab> Date <tab> Other fields.....
847-111-2222 <tab> Omaha <tab> NB <tab> 9/14/2007 <tab>
222-222-3333 <tab> 9/14/2007 <tab>
555-121-1212 <tab> Houston <tab> TX <tab> 9/14/2007 <tab>

Now, if you treat an RTF as a text file, it has only one "plain abfs16f4cf0cb1" after the phone number, so even for the missing line there is only one tab, not 3. This is because in the beginning of the row tabs for each row are defined like this:

tql x90 ql x840 ql....etc...

with "tql" and "tx" tags basically saying where all the tab stops are for that row. So for the row above with missing info, it lists fewer tab stops. So the "date" (and associated garbage) ends up under "City" for this row. All of the "Houston" row's data starts appearing in the sql server output table's 2nd last field, as you might expect.

Any suggestions how to pull this in in SSIS during the transformation? I could deal with it after I pull it in, I still have all the data. I'm thinking the logic to do this could be complicated though. I take the data out of the last two fields of the missing row into some other table, use UPDATES to shift the values 2 fields to the right, and then figure out a way to take the data I just put in a temp table back in, but it all sounds a bit complicated.

Let me know if this makes sense--I've almost got it going, I just need to sort this last bit out.

Thanks,
Kayda