diff options
| author | Narine Kokhlikyan <narine.kokhlikyan@gmail.com> | 2016-01-04 16:14:49 -0800 |
|---|---|---|
| committer | Reynold Xin <rxin@databricks.com> | 2016-01-04 16:14:49 -0800 |
| commit | fdfac22d08fc4fdc640843dd93a29e2ce4aee2ef (patch) | |
| tree | 60f40f7779cd7803848798aa8de37dd55a51fee7 /dev | |
| parent | 34de24abb518e95c4312b77aa107d061ce02c835 (diff) | |
| download | spark-fdfac22d08fc4fdc640843dd93a29e2ce4aee2ef.tar.gz spark-fdfac22d08fc4fdc640843dd93a29e2ce4aee2ef.tar.bz2 spark-fdfac22d08fc4fdc640843dd93a29e2ce4aee2ef.zip | |
[SPARK-12509][SQL] Fixed error messages for DataFrame correlation and covariance
Currently, when we call corr or cov on dataframe with invalid input we see these error messages for both corr and cov:
- "Currently cov supports calculating the covariance between two columns"
- "Covariance calculation for columns with dataType "[DataType Name]" not supported."
I've fixed this issue by passing the function name as an argument. We could also do the input checks separately for each function. I avoided doing that because of code duplication.
Thanks!
Author: Narine Kokhlikyan <narine.kokhlikyan@gmail.com>
Closes #10458 from NarineK/sparksqlstatsmessages.
Diffstat (limited to 'dev')
0 files changed, 0 insertions, 0 deletions