Thanks Jens:
This explanation is very helpful for me. I use relational databases in most of the cases of complex studies. I'm actually needing it right now. And with this explanation I would plan a project starting with the EpiData Classic and maybe when relational and extended checks were implemented in the newer version, I will switch to it importing the database.
Regards,
El 30/01/14 04:57, epidata-list@lists.umanitoba.ca escribió:
Regarding relational databases:
The reality at this point in development is that if you have multiple files in a given project (that is relational databases), then you can do the following: EpiData Classic (3.1): set up a complete relational system, which secures the levels and compliance with rules for 1:1 or 1:many in subfiles.
One example of this is the well known house-person-visit example or the extended example found in a village survey in Burkina Faso. You will find other examples here: http://epidata.dk/examples.php:
EpiData Manager & EntryClient: Each file and subfile will be a separate project (EPX), which you define and enter data into without the appropriate control of 1:1 or 1:many relation. The status for a given project can be followed by reports which will document this . It is the "Count by Id" part of the documentation menu. This "Count by ID" allows the user to select as many fields as desired for identification of a given respondent.
During the coming months the "true" relational setup will be developed. Until then one can do as a user reported last week to have several EntryClients open at the same time and then "manually" change from one project file to the other by selecting on the screen the appropriate project. In each of the projects involved you would define the fields identifying a given observation would be defined as a key, e.g. "id + date" of visit or "id + reason + date of visit + time of day".
EpiData Analysis As well epx and (rec+chk) files can be used for merging relational structures. Although at this point epx files with encryption (password protected) cannot be read into Analysis.
Regarding extended checks Pre-release versions of a scripting language similar to the CHK file contents is now only at the level of "in-house" testing, but this will also come in the test versions later. But assumingly only after relational databases can be setup.
I hope this explains the intentions and limitations.
regards
Jens Lauritsen EpiData Association
Den 30. jan. 2014, epidata-list@lists.umanitoba.ca skrev:
Dear Sir IN relation to the old epidata entry software, As most databases are relational could you clarify what is meant by "in situations with relational databases ."
Dr Thambu David S Professor and Head Medicine Unit 2 Christian Medical College, Vellore 632004, India
"my quote:"
The download page is now suggesting users to get the epx based Manager and EntryClient and only use the old "Classic Epidata Entry" in situations with relational databases and need for extended calculations during entry.
EpiData-list mailing list EpiData-list@lists.umanitoba.ca http://lists.umanitoba.ca/mailman/listinfo/epidata-list