SAP HANA PAL Predict with tree Error exception 73001060

[RR250XC]

Hi Everyone

I am using the PAL and pasting the code from the PAL pdf for Predict with Tree. However, I get the following error:

Could not execute 'CALL "PAL".PAL_DT_SCORING_PROC(PAL_C45_TESTDATA_TBL, #PAL_CONTROL_TBL, ...' in 131 ms 170 µs . SAP DBTech JDBC: [423]: AFL error: [423] _SYS_AFL.AFLPAL:PREDICTWITHDT: [423] (range 3) AFL error exception: exception 73001060: PAL error[73001060]:Internal error. Check trace for details.

All the code is exactly as in the manual except I changed

CREATE TYPE PAL_DT_SCORING_TREEMODEL_T AS TABLE(

"ID" INTEGER,

"TREEMODEL" VARCHAR(5000)

);

to

CREATE TYPE PAL_DT_SCORING_TREEMODEL_T AS TABLE(

"ID" INTEGER,

"TREE_MODEL" VARCHAR(5000)

);

Any help would be greatly appreciated.

[Former Member · Accepted solution]

Hi Rodger,

The current examples for decision trees provided in the guide is kind of inconsistent.

In C4.5 tree, the tree model is given with Json format, and the PMML model table is occupied by decision rules, which cannot be used for scoring. However, the example of Predict with Tree model, it uses PMML model, therefore it leads to the error that tree model cannot be found.

Accordingly, there are two alternative solutions:

1. set parameter IS_OUTPUT_RULES in C4.5 tree be 0, in order to generate PMML model.

2. use Json tree model in Predictive with tree, by setting parameter MODEL_FORMAT be 0, and choose the corresponding model table.

Best regards,

Meilin

[former_member228588]

Hi Rodger.

Can you try changing Integer datatype declaration to Double. I have seen this issue earlier. Please let me know if this is working or not?

Thanks

[RR250XC]

Hi

Still getting the error:

My datatable looks like this:

CREATE TYPE PAL_C45_DATA_T AS TABLE(

"ID" VARCHAR (6),

"SATISFACTION" DOUBLE,

"LAST_EVALUATION" DOUBLE,

"PROJECTS" DOUBLE,

"HOURS" DOUBLE,

"TENURE" DOUBLE,

"ACCIDENT" DOUBLE,

"PROMOTED" DOUBLE,

"POSITION" NVARCHAR (40),

"SALARY" NVARCHAR (6),

"LEFT" DOUBLE

);