I have not linked EpiData to an online database, although we had explored this kind of connection several years ago. Automating the transfer of daily data in EpiData to a central database (MYSQL, Access, or whatever) would probably take more work than the relatively simple task of merging data as EpiData .rec files once the daily files have arrived.
Merging files can be partly automated using EpiC (download from epidata.dk) and a simple WINDOWS/DOS batch file. In a Unix/Linus environment this could also be automated easily with GREP as the file structures will not change and can be stripped from the daily files easily.
Files can be uploaded to a designated folder, merged into the master database and individually copied to an archive folder. This gives you 2 copies of all of the data for safety and in case the merge does not go well.
The advantage of keeping everything in EpiData is simplicity. All you have to program is a couple of batch files that could be as simple as the following (I haven't tried this recently, but have done this before) and then set up the folders appropriately.
* add1.bat * append a single daily file to the master file copy %1 archive/%1 rename master.rec tempmaster.rec epic append tempmaster.rec %1 master.rec * add in some check for error after epic erase tempmaster.rec
* addall.bat * append all daily files copy master.rec archive/master.rec for %%f in *.rec do add1 %%f
Jamie Ottawa
On 2010-03-25, at 6:48 AM, epidata-list@lists.umanitoba.ca wrote:
I am to start up a large study where several subinvestigators will have to enter data from there own computer on a daily basis. I think it would be an advantage to the data management, if we did not have to merge all the datafiles manually, but instead enter them into an online database after each entry session.