Hi Ayton Glad to hear that you have built a mart based on our solution.
On the tables being empty issue, I am sending you (offline to this list) two sql dumps for you to create a normalized table structure and to create the mart based on it. These are the data dumps from the example we have given. If you have any other issue please feel free to contact us. Gnani ________________________________ From: Ayton Meintjes [mailto:[email protected]] Sent: September 1, 2009 9:41 AM To: Saravanamuttu Gnaneshan Cc: [email protected]; Christina Yung Subject: Re: FW: [mart-dev] Dynamic attribute set Hi Gnani, Christina Thanks for the quick response (and for providing an example), this is exactly what I was looking for. Got it working on a test dataset already. regards Ayton Having some issues still though. The structure is being created but the tables are empty 2009/8/28 Saravanamuttu Gnaneshan <[email protected]<mailto:[email protected]>> Hello Ayton We have now built a test mart to satisfy your requirements with some fake data as shown in your mail. This mart can be browsed from this url: http://dcc-dev.res.oicr.on.ca:9030/biomart/martview/3e46bb17a7de44e0df5e2e8deab7715c. If the above link expires by the time you see this mail please use this url: http://dcc-dev.res.oicr.on.ca:9030 This mart was built as explained in Christina's email and if you need to see the XML files created by the martbuilder and marteditor please email me offline (not through the mart-dev list). If you have any queries on this please feel free to contact us. Best wishes Gnani Saravanamuttu Gnaneshan Scientist, Computational Biologist Ontario Institute for Cancer Research MaRS Centre, South Tower 101 College Street, Suite 800 Toronto, Ontario, Canada M5G 0A3 Tel: 647-260-7990 Toll-free: 1-866-678-6427 www.oicr.on.ca<http://www.oicr.on.ca> This message and any attachments may contain confidential and/or privileged information for the sole use of the intended recipient. Any review or distribution by anyone other than the person for whom it was originally intended is strictly prohibited. If you have received this message in error, please contact the sender and delete all copies. Opinions, conclusions or other information contained in this message may not be that of the organization. ________________________________ From: [email protected]<mailto:[email protected]> [mailto:[email protected]<mailto:[email protected]>] On Behalf Of Christina Yung Sent: August 27, 2009 9:53 AM To: Ayton Meintjes Cc: [email protected]<mailto:[email protected]> Subject: RE: [mart-dev] Dynamic attribute set Hi Ayton: I think a mart with partititioned tables in the following structure will serve your needs: Marker [id, rsid, allele_A, alleleB] CEU_AlleleFrequency [fk_marker_id, CEU_freq] JPT_AlleleFrequency [fk_marker_id, JPT_freq] YRI_AlleleFrequency [fk_marker_id, YRI_freq] In this mart, Marker is the main table where *_AlleleFrequency are the dimension tables. This structure allows the addition of new populations as dimension tables. Keep in mind that when multiple *_freq attributes are selected, multiple tables are joined. Hence, scalability may become an issue if the number of populations is large. To arrive at this mart, you may wish to use MartBuilder. First, specify Marker as the main table. MartBuilder will merge Population and AlleleFrequency into one dimension table [id, fk_population_id, fk_marker_id, freq, population_name]. The next step is to partition this dimension table by either fk_population_id or population_name. Hope this helps. Christina ________________________________ From: [email protected]<mailto:[email protected]> [mailto:[email protected]<mailto:[email protected]>] On Behalf Of Ayton Meintjes Sent: Wednesday, August 26, 2009 11:37 AM To: [email protected]<mailto:[email protected]> Subject: [mart-dev] Dynamic attribute set I have the following three tables in a custom database: Population [id, name] Marker [id, rsid, allele_A, allele_B] AlleleFrequency [id, fk_population_id, fk_marker_id, freq] How would I go about creating a mart that the user can query for markers and get results in the form: rsid | CEU_freq | JPT_freq | YRI_freq rs12 | 0.05 | 0.55 | 0.75 rs34 | 0.95 | 0.65 | 0.05 In other words, the number of attributes available depends on how many unique populations there are? The automate push action and partitioning seem to suggest this sort of structure might be possible. Thanks in advance
