Inter-table relationships

[soup_dragon]

I'm looking at a pair of imported text files to mine information about shared rows (keyed on a shared field in each row). When I define a relationship and update the view, then filter on related records I see on a single entry - and I know for sure that there are many shared rows.

Am I doing something wrong or are there known bugs with this feature?

thanks

SD

[Ben]

A few questions:

1) Are that there are no whitespace considerations, such as leading tabs or spaces?

2) Certain types of relationships might need a little help, i.e. if you have a date inside a character field in one table, and the data in a date-typed field in the second table, you might have to indicate in the relationship setup how to match these values up.

If you'd like, I'd be willing to take a look at the data myself-- just post a link.

All the best,
Ben

[soup_dragon]

Hi Ben,

Thanks for the offer of looking at these files. I can't see any whitespace issues with the data columns.

The two example files are at:

http://www.sanger.ac.uk/Users/avc/STRATA/test1.txt
http://www.sanger.ac.uk/Users/avc/STRATA/test2.txt

I'm trying to define the relation ship on column 2 of each file

thanks

SD

[Ken]

I took a look at those two files and grouped them (based on field2) to count any multiples, but it appears that each row is unique in its given file. Could you give me a value in field2 from test1.txt that should have multiple related rows in field 2 in test2.txt?

We'll eventually get down to the bottom of this.

Thanks,
ken

[soup_dragon]

Ah, that may be the source of the confusion. I'm looking to extract all rows from both files that share a field2 value - not a single-to-many relationship. I'm looking for the intersection of the two files based on field2

SD

[Ben]

Hello,

This is possible. Use the query builder (New->Query) and add your tables to the top portion. Next, insert the fields you want in your output file. After you set up a relationship between the two tables, you can click on the line to adjust the join type (INNER, LEFT OUTER, etc).

The regular relationship manger is also very useful for on-the-fly aggregate calculations, but in this case, the Query Builder implements the style of join that you are interested in.

Hope this helps,
Ben

[soup_dragon]

Thanks Ben - I have this working via the query interface as you explained. Strata really is very cool so far in my experience. I'm particularly impressed with the way it seems to out to "do the right thing" most of the time. It certainly has made my analysis across a series of data sets much easier to handle - particularly because the data models are not all alike, some are too big for XL and a full blown database solution is too heavy weight for the scale of the analysis.

All in all a very nice tool to have around!

I guess my main observation is that the interface does not clearly expose some of the powerful features and I have had trouble finding examples to learn more about them.

I guess that documentation will come in time though.

thanks

SD.

[Ken]

Thanks for the observation and comments! We are busily working on getting documentation sorted out and examples are close behind. We hope to have a lot more available in the coming weeks. Feel free to ask any questions you have in the meantime.

All the best,
ken