on 07-26-2013 7:19 PM
Hi Community,
We're trying to create an R procedure that returns some error statistics from a prediction that was successfully made from an R model in another script. Unfortunately, when we try to run this procedure, it gives an error:
--
SAP DBTech JDBC: [2048]: column store error: [2048] LIHUA.GET_SUM: line 5 col 1 (at pos 94): GenericFailure exception: column store error: search table error: [34086] Convert data error: Convert Data from R to NewDB failed.;Column 'NAME' not computed from R.
--
The script worked fine when run in R and showed the results in a table, as expected. If I remove the column 'NAME' from the data, it gives the same error about the new first column.
The R code that puts the results into a table is as follows (ER and MSE are R functions defined earlier in the script, line breaks were added for readability):
result<-data.frame(Name=c("ALL","STEP"),
Interval_Error=c(Inter(pre$PRE_ALL,pre$PRE_ALL_LOWER,pre$PRE_ALL_UPPER),Inter(pre$PRE_STEP,pre$PRE_STEP_LOWER,pre$PRE_STEP_UPPER)),
Tolerance_Error=c(ER(pre$PRE_ALL,pre$SALES,3),ER(pre$PRE_STEP,pre$SALES,3)),
MSE=c(MSE(pre$PRE_ALL,pre$SALES),MSE(pre$PRE_STEP,pre$SALES)))
The table being written to is this format:
DROP TABLE SUMMARY_ERROR;
CREATE COLUMN TABLE SUMMARY_ERROR (
Name VARCHAR,
Interval_Error FLOAT,
Tolerance_Error FLOAT,
MSE FLOAT
);
Has anyone experienced anything similar to this while running R scripts in HANA?
Thanks,
Dave
The issue turned out to be one of case; the columns being returned from R needed to have the same name as the columns in HANA, and HANA column names are all uppercase.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi David,
Please check below link
https://saptechedhandson.sap.com/demo.sap.com~vi~web/content/CD166.pdf
it may help you or provide other source of information.
Best Regards,
Atul
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
User | Count |
---|---|
101 | |
13 | |
13 | |
11 | |
11 | |
7 | |
6 | |
5 | |
4 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.