如何检测数据库中的变化？

Question

This maybe the problem of workflow in my organization, but I would like to explore whether this can be fixed/alleviated technically.

I work in a healthcare organization that data about a certain illness is constantly retrieved for compilation of reports. I'm somewhere in a middleman role, the IT team export data for us, and we validate the data. Since we don't have access right to the raw data, so we can only voice out the issue of the data to IT team, they work on the amendment according, if they REALLY wiLL, the amendment maybe re-retrieval of data, or maybe asking the nurses to enter the missing data again.

I communicate with IT team primarily using email, since we don't share any data server, so data is usually packed in excel and sent by email. Problem arises when there are several rounds of data checking and amendment, when the new data arrived, I can only believe that the amendment is done and I need to check everything again, and when there are 2-3 versions of data coming to and fro, things can be very messy. (did the 4th amendment also incorporate 1st, 2nd, 3rd amendments also? Or other than my request, new amendment is included also?)

I start using GIT to detect and store changes of my scripts, I wonder if I can use git to rapidly check the database, I can export the ~50000 rows, ~100 columns table into csv, is it feasible?

What I want to do is to rapidly check the data (ie I said row XYZ need to change column ABC, did IT team really fix it? Did IT team fix other things that I don't knwo?) Any suggestions? Thanks!

Answer 1

You seem to be attacking the problem in the wrong place.

The database itself ought to be throwing bad data right back at the person who entered it. Keeping bad data out of the database is a lot easier and cheaper than fixing it after it gets in. (Although I guess that might put you out of a job.)

Having said that, you can use any version control system to follow changes, although deleted rows might cause you some problems.