Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
I expose an example,I think is better to can answer my question.
Suppose a Table of pharmacy
patient id drug date_pharmacy center service assistance
1 A 20150101 1 100 15H100
1 B 20150203 1 201 15H320
2 A 20150210 1 100 15H330
2 A 20150705 2 500 15H403
2 C 20150705 2 200 15H403
3 A 20140312 2 100 14H342
4 B 20140123 1 200 14H001
other of laboratory
patient id center service assistance test date_lab
1 1 200 15H100 3 20150101
1 3 423 15H100 2 20150102
1 2 200 15H191 1 20150124
3 1 100 14H342 3 20150801
other Radiology
patient id center service assistance RX date_rad
1 1 500 15H100 123 20150101
6 3 100 15H111 23 20150223
1 2 100 15H400 123 20150227
if I concatenate.
patient id drug date_pharmacy center service assistance test date_lab RX date_rad
1 A 20150101 1 100 15H100
1 B 20150203 1 201 15H320
2 A 20150210 1 100 15H330
2 A 20150705 2 500 15H403
2 C 20150705 2 200 15H403
3 A 20140312 2 100 14H342
4 B 20140123 1 200 14H001
1 1 200 15H100 3 20150101
1 3 423 15H100 2 20150102
1 2 200 15H191 1 20150124
3 1 100 14H342 3 20150801
1 1 500 15H100 123 20150101
6 3 100 15H111 23 20150223
1 2 100 15H400 123 20150227
I have a lot of Fact tables, and I need to can answer questions beetween fact tables, and I don't know if concatenation is the best way or is best link tables.
Can I answer, with concatenation, the patients of drug A , that they have a test type 3?
Can I answer,with concatenation, the assitances of laboratory type=3 , that they have a Radiology test?
It's possible to make two or more link tables?.
link tables for:
patient id
medical assistance id
center,
healthcare_service,
.
Thank you very much.
Unfortunately there isn't a best way - it will be depend on many factors which approach is suitable for you but this Concatenate vs Link Table and this Get started with developing qlik datamodels will be helpful for you.
- Marcus
Since the tables have a lot of fields in common my first thought is to see if the tables can be concatenated. I would rename the date fields so you have only one date field and I'd add a source field to store where the data came from.
Can I answer, with concatenation, the patients of drug A , that they have a test type 3?
Yes. The easiest way is to first select drug A, then select the possible patient_id values. Next unselect drug A. Finally check if any selected patient id's have test type 3. If test type 3 is a possible value there are patients with drug A and test type 3.
It's also possible to create set analysis expressions that do this kind of filtering in an expression. Something like:
count({<drug={'A'},[patient id]=P({<drug=,[test type]={3}>}[patient id])>} [patient id])
Can I answer,with concatenation, the assitances of laboratory type=3 , that they have a Radiology test?
Yes, using the same procedure as above.