Not reading new Data

[datawrangler]

Good grief.. suggestions?
thanks for your responses to my posts. i was doing something goofy with my suppress, i guess... because it's working fine now.

Here's my latest issue: Now, I have everything built and I've started doing my daily updates, bringing in new daily sales data. I must have messed something up because now, it's not bringing in the most current data. I had some colums mismatched last week, so I fixed that and unchecked the 'Disconnect from data source' box in my preferences, and my build worked fine. Now it's happening again this morning. I know my colums match, so I unchecked the Disconnect from data source, and re-built... no new data. I must have messed with my preferences or something. Any ideas? Again, thanks for the responses.

 

[DoubleD]

1. Make a backup of your MDL.
2. Clean out all your categories except for custom categories.
3. Save your MDL.
4. Close Transformer.
5. Build your Cube.

Ensure you are looking at the most recent version of the MDC.

I am what I am based on the decisions I have made.

DoubleD

 

[datawrangler]

I know I must sound like an idiot, but what's the best way to do this? Right in the dimension veiwer? I have quite a few custom categories for grouping.

Thanks again...

 

[DoubleD]

If you have a lot of custom categories, then the Dimension Viewer is definitely your best bet.

I don't know whether you have both a development and production environment, but if you do, ensure you're grabbing the data from the appropriate environment. Many times I've thought I wasn't getting the new data, only to find out I was still pointing to development, or vice versa.

I am what I am based on the decisions I have made.

DoubleD

 

[datawrangler]

Well, the whole thing is ridiculous right now. It's based off this huge msaccess table until we can get a real data warehouse built. everything seemed to be going fine at first. I know that I'm accessing the right table (there's only one), and I know the table is up to date. Transformer is also reading the proper number of records during the first step of the build, it's just that the data is not showing in the final the mdc. For instance, I have data from all years. I replace all of 2004 data every day (in my source). When I build the cube, all the old data is there and 2004 is blank (although it's definitely in the data source). Last week I was able to match my columns and it fixed the problem... I'll try the category thing.

 

[DoubleD]

Are you sure you're not excluding any dimensions or measures from your cube? Check the properties of your cube in the MDL to verify that you aren't building specific views or excluding any data. Also ensure the data is falling into the appropriate custom categories.

I am what I am based on the decisions I have made.

DoubleD

 

[datawrangler]

you didn't know what you were getting in to when you decided to help me.

Well, I only have .pyi files. I can't open .mdl files because it doesn't like some of my custom categories. I called Cognos and asked them how to fix these "Category associations that i had evidenlty deleted" and they said not to bother trying to open .mdl as it saves in layers. I dunno, maybe I'm not supposed to open the .mdl.

My custom categories were working fine in the .pyi and final .mdc's. I'm not excluding anything out of the ordinary. I recall messing with the .mdl (.pyi) preferences last week and that's when the trouble started.

 

[DoubleD]

Datawrangler,
Without knowing what things you might have changed, I'm not sure I'm going to be of much assistance. Here are the last of my suggestions, then I'm at wits end.
1. If you changed your default date format in your preferences, your PYI may not be recognizing your dates properly anymore, this would explain why they wouldn't fall into the 2004 category.
2. Your PYI may be corrupt, in that instance, all you can do is build a new one.

I would suggest working with MDL files instead of PYI files. PYI files have a tendency to get corrupted. I've never had an MDL get corrupt.

Good luck with this, I hope you find a solution. If you've got anymore valuable clues, let me know.

If you can't figure out anything, and your data isn't sensitive, you could send me a portion of your data, along with your PYI and I could take a look at it.

I am what I am based on the decisions I have made.

DoubleD

 

[datawrangler]

hey, DD

I figured it out. The launch dates that I added to my table weren't updated correctly, the 2004 had launch dates of 0. I fixed that and all is well now. About MDL vs. PYI. it freaks out when I try to save it as an MDL. It locks up and says I've apparently deleted some category associations somewhere, the the whole thing corrupts and dies. It works fine as a PYI. And there's no way I'm rebuilding this mess! GODFORBID...

Do you know anything about dates? My dates seem to be working properly, except when I get down to the day level.
For Instance, I show nothing for the first week of Sept 2003 and I know we had sales. That's throwing my MTD comparison off. I'm sure it's the way I split the weeks or something, but I can't figure it out. Suggestions on either of these? Thanks for all your help.

 

[DoubleD]

If you're running dates at the Week level, it depends on what you're using for a starting date. For instance, the first three days of 2004 might be thrown out because that's Thurs, Fri, Sat.
If you set your dates in your model as Lunar, they will go by full weeks, but this will make your Monthly totals be 28 or 25 day groupings instead of matching the calendar.

I am what I am based on the decisions I have made.

DoubleD