Create Longitudinal project in REDCap

[wibeasley]

@thomasnwilson, thanks for agreeing to do this. As of right now, I think this project needs just four columns.

1. record_id: autonumber field (recommended by @thomasnwilson )
2. database: a text field
3. table: a text field
4. probe_date: a date and time field
5. row_count: an integer field

Please call my house if you have questions.

[thomasnwilson]

I already created a shell project called Database Management. I have given you @wibeasley and David @bard1536 access to it. I gave you access to the user rights so you can adjust your own rights to what is appropriate. If you would like me to change the name of the project let me know.

[thomasnwilson]

@wibeasley

The four variables have been added. REDCap automatically creates a variable named record_id. So, right now your project has 5 variables (record_id plus the 4 you requested). In your R code, you may want to have some sort of autonumber to fill in the values for record_id. I could make the record_id be the database variable, but this field has to have a unique value.

FYI, probe_date is set to a M-D-Y H:M:S. If you need it in a different format, let me know.

I will leave this project in the development stage for now. When you're ready, you can move it to production, or I can. Just let me know.

[wibeasley]

@thomasnwilson, I think we'll eventually want to change the name from 'Database Management', at least because we'll have two sister REDCap projects where that name could apply. But this is fine for now. Is it pretty easy to change later? I assume the project _id won't change, and that's what the backup tool needs to know (not the name that humans see in the webpage).

@bard1536, the following is more for your benefit.
The first project is this one, that longitudinally tracks the record count of each table in each database. If a database has 5 tables, after six days there will be 30 records (=5*6).

The second project is more like demographic information, where there's just one record per database. Currently, I'm thinking:

1. database a text field for the name
2. server a text field for the name of the server
3. pi_name the principle investigator
4. project_manager_name the person who actually knows what's going on
5. server_type right now, it we'll support only SQL Server. Maybe Access or MySQL later. The backup tool (which is probably written in R) will need to know, because they'll be handled differently
6. active a boolean field for whether the tool should still be backing up the database

[thomasnwilson]

@wibeasley It is very easy to change the name later. Only takes a few seconds. I am pretty confident that the project_id will stay the same. If you what you want to name it, I can change it now.

[wibeasley]

Thanks @thomasnwilson, it looks great. Thanks for setting it up, and saving me a lot of time figure out the system. I promise to learn & do it myself for the parent project.

I did change the date format to ISO-8601, for several reasons. Partly because R handles it natively. Partly because I like it better. Partly because it gives me an excuse to reference an XKCD cartoon.

[thomasnwilson]

@wibeasley All good reasons.

[wibeasley]

For the record, I changed it from 'Database Management' to 'Database Management Log'. You're right that the project_id didn't change.