cancel
Showing results for 
Search instead for 
Did you mean: 

PA Error: SAP DBTech JDBC [259] invalid table name PROCEDURE_SIGNATURE_T

Former Member
0 Kudos

Hi Experts,

I am getting this error message about PROCEDURE_SIGNATURE_T table being an invalid table name while using the most of PAL functions through Expert Analytics (Some functions work without problem like Decision tree 4.5).  We 're on HANA revision 93, PA version 2.3.

I saw SAP note 2128915 that recommend revision 96 (93 is for test and evaluation only) and in other SCN question that hana revision 96 + PA 2.2 + APL 2.2 will be OK:

I want to know where installing this APL , is it a component of hana database like PAL ?  and which version will be compatible with PA 2.3  ?



Best regards,

Abdess

Accepted Solutions (1)

Accepted Solutions (1)

achab
Product and Topic Expert
Product and Topic Expert
0 Kudos

Please also refer to post here:

Best regards

Antoine

Former Member
0 Kudos

Hi antoine !

For this time when I run the hana auto classification algorithm i found this error on the screenshot :

Thanks,

Abdess

achab
Product and Topic Expert
Product and Topic Expert
0 Kudos

Please check the hana rights required by APL (from the doc link I shared earlier on)

Former Member
0 Kudos

I applied what is in the doc, but i still have the following error :

when i skip the culumn indicated in the screenshot, the error goes to the next column.

Best regards,

achab
Product and Topic Expert
Product and Topic Expert
0 Kudos

Hi,

the good news is that the error does not seem to be related to APL. Can you consume the HANA view from HANA Studio and view the related data ?

Best

Antoine

PPaolo
Advisor
Advisor
0 Kudos

Hello,

another point to test is the name of the columns in HANA.

Can you check if you are able to read columns which don't contain spacesint their name?

The attribute you are not able to read is 'Number of children': we have sometimes seen unexpected behaviours when there were spaces in column names.

Regards

PPaolo

achab
Product and Topic Expert
Product and Topic Expert
0 Kudos

Hi Abdess,

Have you been able to resolve your problem?

Thanks & regards

Antoine

Former Member
0 Kudos

Hi !

I deleted the spaces on the columns name but I still have the same problem,

I can view correctly the data from the hana studio. See screenshot.


Best regards !

Abdess

achab
Product and Topic Expert
Product and Topic Expert
0 Kudos

Hi,

Are you able to run the first node in Expert Analytics (the dataset one) and get the values back?

Thanks & regards

Antoine

Former Member
0 Kudos

If you are talking about the node "prepare" then yes.

Thanks,

achab
Product and Topic Expert
Product and Topic Expert
0 Kudos

No I am thinking of the very first node in your process chain in the Predict room.

Former Member
0 Kudos

Hello !

I created a new dataset for test without sapces in the names of culumns. this time i have the same error but with a culumn name that not exist in the dataset.

When i use the hana C4.5 algorithm it works but when I add hana model statistics I find the same error.

Thanks,

Former Member
0 Kudos

Yes, I get the values back. No problem in this node.

achab
Product and Topic Expert
Product and Topic Expert
0 Kudos

Have you assigned the proper HANA security to your APL user?

Thanks & regards

Antoine

Former Member
0 Kudos

Yes, of course.

Regards,

Abdess

achab
Product and Topic Expert
Product and Topic Expert
0 Kudos

It would be great if you can retrieve the two log files corresponding to the tests performed on the HANA view with spaces and the HANA view without spaces so that we can take a look.

Log names start with Tracelog and are found here: C:\Users\<your Windows user>\AppData\Local\Temp\sappa\logs.

Please close Expert Analytics and replay the failing workflow to generate "clean" log fileq.

Sorry for all the inconvenience faced.

Best regards

Antoine

Former Member
0 Kudos

Hi Antoine !

Thanks a lot for your help.

Please find attached the tracelog when applying hana auto classification algorithm with a dataset that hasn't spaces in its columns names.

Regards,

achab
Product and Topic Expert
Product and Topic Expert
0 Kudos

In fact, it is even documented as a known restriction in the APL user guide on page 200.

https://websmp206.sap-ag.de/~sapidb/012002523100013455472015E/pa23_hana_apl_user_en.pdf

HANA Table and Column Names

When defining and using table types with APL, blanks are not supported in column names: The AFL wrapper generation fails when a table type contains column names with spaces. This results in the following error message: "AFL error exception: AFL error: Internal error : generating parameter string failed."

achab
Product and Topic Expert
Product and Topic Expert
0 Kudos

Hi,

Can you share a data snippet with me ? I would need a table that I can export on my HANA system.

Could you please try importing the attached data table and see if it works? this is a table with CENSUS table. It is working for me.

Best regards,

Antoine

achab
Product and Topic Expert
Product and Topic Expert
0 Kudos

Also, please pay attention to the following release restriction note: http://service.sap.com/sap/support/notes/2178039

My current understanding is that the restriction also applies to APL (as APL relies on AFL) in addition to PAL. 

You need to check the data types for all the columns of the tables involved in your HANA views, to make sure they comply with the below list 

Best regards

Antoine

Unsupported columns in table types:

    • Processing of data may fail while using Predictive Analysis Library (PAL) components in Expert Analytics since the Application Function Library (AFL) does not support certain columns in table types.
    • The following are the supported CSTYPE (column store type)/SQLTYPE/DIMENSION combinations from AFL:
      • "INT" / "INTEGER"
      • "FIXED8_19_0" / "BIGINT" ("FIXED8_19_0" means fixed with 8 bytes and 19 digits)
      • "DOUBLE" / "DOUBLE"
      • "STRING" / "CLOB"
      • "STRING" / "NCLOB"
      • "STRING" / "VARCHAR" / ...
      • "STRING" / "NVARCHAR" / ...
      • "DAYDATE" / "DATE"
      • "SECONDTIME" / "TIME"
      • "LONGDATE" / "TIMESTAMP"
      • "SECONDDATE" / "SECONDDATE"
Former Member
0 Kudos

Hi Antoine !

Finally, I resolved my problem. It was about two things :

First, the columns should not have labels(descriptions). When I filled columns labels, Expert Analytics returned the error : " ... attribute not found in column table : 'column name' ..." (like if this lablel replace the ID of the column).

Second, for the error  "... Invalid column name:ApplyIn ..." I added the field "ApplyIn VARCHAR(100)" to the table type "PROFITCURVE_T" and It works fine.(https://websmp204.sap-ag.de/~sapidb/012002523100013455472015E/pa23_hana_apl_user_en.pdf Section 3.3.4)

Thanks for your help.

Abdess,

achab
Product and Topic Expert
Product and Topic Expert
0 Kudos

Hi,

That's good to know!

Enjoy APL!

Antoine

Answers (1)

Answers (1)

achab
Product and Topic Expert
Product and Topic Expert
0 Kudos

Hi Abdess,

I am not very clear about your problem statement.

We have to distinguish two different predictive libraries in SAP HANA: APL and PAL.

APL is the implementation of the Automated Analytic powerful algorithms into SAP HANA. All algorithms with prefix HANA Auto make use of APL.

PAL is a comprehensive data science algorithm library, providing among others decision tree C4.5 classic algorithm for instance like in your example.

If I am not mistaken, both of these predictive libraries require a separate installation and configuration of top of the "base" SAP HANA installation.

A few questions to better understand where you stand:

- Have you installed both APL & PAL?

- Have you checked that the configuration is successful?

For APL, this means following the guide here: https://websmp106.sap-ag.de/~sapidb/012002523100013455472015E/pa23_hana_apl_user_en.pdf from page 9 to page 15.

For PAL, the online resource available right now is this: http://help.sap.com/hana/SAP_HANA_Predictive_Analysis_Library_PAL_en.pdf that applies to HANA SPS10.

There is a difference between the error message that you mention : "I am getting this error message about PROCEDURE_SIGNATURE_T table being an invalid table name" and the screenshot. The algorithm you used at least for the screenshot is a APL one, this would lead to suspecting a APL configuration or installation one.


Can you kindly provide the full logs? they are available here: C:\Users\<your Windows user>\AppData\Local\Temp\sappa\logs, take the more recent one, the name starts with TraceLog.


The PA 2.3 PAM claims support of SAP HANA SPS9 based on revision 96 which is the revision that is recommended by SAP for productive usage. With PA 2.3 you can use either APL 2.2 or APL 2.3.

Good luck,

Best regards


Antoine

Former Member
0 Kudos

Hello Antoine !

We upgraded hana database to revision 97, we installed PAL and APL successfully. But we have 2 problems :

For APL algorithms like HANA Auto Classification: Invalid table name PROCEDURE_SIGNATURE_T. See the screenshot :

For PAL algorithms like HANA C4.5 : Insufficient privilege : Not authorised.

I don't know exactly what privileges should i affect to the user so that the algorithms of PAL and APL work.

If you have a documentation datailed the steps of connection of expert analytics and hana database, it will be perfect.

I hope that i detailed my problem well.

Thanks a lot for your help.

achab
Product and Topic Expert
Product and Topic Expert
0 Kudos

Hi Abdess,

APL problem: see my answer below.

Please refer to Jayanta Roy post here: Solve the APL configuration issue for Expert

PAL problem: please refer to the user guide as I mentioned. http://help.sap.com/hana/SAP_HANA_Predictive_Analysis_Library_PAL_en.pdf

Here is what I see in the section related to Security

Cheers

Antoine

achab
Product and Topic Expert
Product and Topic Expert
0 Kudos

Additional info that can be useful for you.

http://service.sap.com/sap/support/notes/2178039

  • Unsupported columns in table types:
    • Processing of data may fail while using Predictive Analysis Library (PAL) components in Expert Analytics since the Application Function Library (AFL) does not support certain columns in table types.
    • The following are the supported CSTYPE (column store type)/SQLTYPE/DIMENSION combinations from AFL:
      • "INT" / "INTEGER"
      • "FIXED8_19_0" / "BIGINT" ("FIXED8_19_0" means fixed with 8 bytes and 19 digits)
      • "DOUBLE" / "DOUBLE"
      • "STRING" / "CLOB"
      • "STRING" / "NCLOB"
      • "STRING" / "VARCHAR" / ...
      • "STRING" / "NVARCHAR" / ...
      • "DAYDATE" / "DATE"
      • "SECONDTIME" / "TIME"
      • "LONGDATE" / "TIMESTAMP"
      • "SECONDDATE" / "SECONDDATE"

Note: Not all of these columns may be supported with PAL. For more information, see the PAL documentation at http://help.sap.com/hana_one/ -> SAP HANA Predictive Analysis Library (PAL) Reference.