If there are no cycles (i.e., release ranges) in the database for the
given project - which happens in the ML test, for example -, the
subsetting of the result throws an error.
By, returning an empty data.frame instead, everything is fine. The user
only needs to check for "nrow(res) > 0" if they rely on the result.
Signed-off-by: Claus Hunsen <hunsen@xxxxxxxxxxxxxxxxx>
---
codeface/R/query.r | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/codeface/R/query.r b/codeface/R/query.r
index 4911136..72c8504 100644
--- a/codeface/R/query.r
+++ b/codeface/R/query.r
@@ -137,9 +137,11 @@ get.cycles.con <- function(con, pid, boundaries=FALSE,
allow.empty.ranges=FALSE)
column.selection <- c("range.id", "date.start", "date.end", "cycle")
}
- res <- res[, column.selection]
- res$date.start <- ymd_hms(res$date.start, quiet=TRUE)
- res$date.end <- ymd_hms(res$date.end, quiet=TRUE)
+ if (nrow(res) > 0) {
+ res <- res[, column.selection]
+ res$date.start <- ymd_hms(res$date.start, quiet=TRUE)
+ res$date.end <- ymd_hms(res$date.end, quiet=TRUE)
+ }
return(res)
}
--
2.10.0
