Handle PHP backend databases without full datasets better

Many thanks to Andreas Voss for uncovering this issue!
ctrl-labs · Apr 6, 2020 · 3869435 · 3869435
1 parent fff8dfd
commit 3869435
Show file tree

Hide file tree

Showing 3 changed files with 11 additions and 0 deletions.
diff --git a/readme.md b/readme.md
@@ -113,6 +113,7 @@ Dilek Akkus ·
 [Nikoletta Symeonidou](https://www.sowi.uni-mannheim.de/en/erdfelder/team/academic-staff-members/symeonidou-nikoletta/) ·
 [Hendrik Singmann](http://singmann.org) ·
 [Kevin Tiede](https://www.spds.uni-konstanz.de/kevin-tiede) ·
+[Andreas Voss](https://www.psychologie.uni-heidelberg.de/ae/meth/) ·
 [Andrew White](https://github.com/shiroandy) ·
 [Malte Zimdahl](http://psycho3.uni-mannheim.de/Personen/Malte%20Zimdahl,%20M.Sc./)
 

diff --git a/utilities/SQLite import app/server.r b/utilities/SQLite import app/server.r
@@ -110,6 +110,11 @@ processData <- function(database) {
       ) %>%
       ungroup() %>%
       select(-id) -> d.full
+  } else {
+    # If there are no full datasets, start from an entirely empty df
+    # in order to avoid introducing unwanted columns into the following
+    # merge steps.
+    d.full <- tibble()
   }
 
   d %>%

diff --git a/utilities/SQLite import.Rmd b/utilities/SQLite import.Rmd
@@ -127,6 +127,11 @@ if (nrow(d.full) > 0) {
     ) %>%
     ungroup() %>%
     select(-id) -> d.full
+} else {
+  # If there are no full datasets, start from an entirely empty df
+  # in order to avoid introducing unwanted columns into the following
+  # merge steps.
+  d.full <- tibble()
 }
 ```