Closed klapaukh closed 2 years ago
Merging #148 into master will decrease coverage by
3.08%
. The diff coverage is0%
.
@@ Coverage Diff @@
## master #148 +/- ##
==========================================
- Coverage 75.65% 72.56% -3.09%
==========================================
Files 14 15 +1
Lines 1269 1323 +54
==========================================
Hits 960 960
- Misses 309 363 +54
Impacted Files | Coverage Δ | |
---|---|---|
R/dbConnection.R | 0% <0%> (ø) |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update a21af68...a5995eb. Read the comment docs.
Is the AppVeyor failure specific to this PR, or a long-standing problem?
Seems to have been failing for a while, including all of #145. From #135 it may be caused by importing dplyr?
It is really great to see you guys starting to implement the database approach. Although I don't know exactly what you are planning to do, cleanEHR now has been branded as a more generic data cleaning package, at least this is what we said in the JOSS paper, which is still under review. Moving from ccdata to database fits the purpose well. As you named the branch as "pipeline", I wonder if this is the data processing pipeline only for CCHIC, e.g. import XML files from hospitals to the database, you might want to consider to move it to a separate package, at least until the JOSS paper is published.
@sinanshi - the database importer is done in a different package, this is only to read from a database and generate a ccdata object.
@dpshelio that's great.
This basically allows you to create ccd objects not from the XML or the RData, but rather than the database using the following: con <- connect(username="something", database="something") ccd <- table.to.ccdata(exportData(con) %>% collect, metadata(connection = con))
The only real changes are in R/dbConnection.R [and the README but those are trivial]