No catalog data extracted for HTTP/HTTPS import action

[laurent_burtaire]

Hello,

we are implementing SRM UI Add-On with software components:

• SRM 7.0,
• SRM_SERVER 714 SP 2,
• SAP_BASIS 7.50 SP 3,
• SRMNXP01 150 SP02,

We succeeded in SRM Product Master Data extraction and product data indexation with TREX.

The Cross-Catalog Search works correctly.

Nevertheless, with catalog extraction in HTTP/HTTPS import, no data are extracted.

Indeed, the call:

• CALL METHOD lv_res_client->response->get_cdata
  RECEIVING
  data = lv_body_string.

in lv_body_string, there is no items detail, so process cannot carry on.

Moreover, we do not have any log after background process with report /SRMNXP/CATALOG_SCHEDULER.

Does anyone have some clues regarding this extraction process?

Regards.

Laurent.

[laurent_burtaire]

Hello,

nobody knows?

Regards.

Laurent.

[laurent_burtaire · Accepted solution]

Hello,

i have the final answer...

Regarding catalog data extraction from Ariba in JSON format, parameters to configure requests are available in the integration guide for Ariba Procurement Content:

- "URL" > correct URL is https://s1-eu.ariba. com/Buyer/downloadJson/<ANID> where <ANID> is the Ariba Network ID value for Ariba Procurement Content (this URL will never be the same as the one for Punch-Out catalog access),

- "UserId" and "password" are used to connect to APC for data extraction. Nevertheless, those two mandatory parameters do not exist for Punch-Out integration with Ariba catalog. This means they do not make part of Web Service catalog call structure.

Moreover, in SAP standard program /SRMNXP/CATALOG_EXTRACTION (transaction /SRMNXP/CAT01), user and password parameters are hard-coded (USERNAME and PASSWORD), so, it cannot work if you populate selection fields:

The only workaround is to add parameters "UserId"' and "password" in the Ariba catalog call structure Web Service for Punch-Out.

Be carefull, the HTTP request is case sensitive for its parameters. So, take care of the upper or lower case for each parameters given in the APC integration guide (otherwise, you will get HTTP 401 error) !

For any other Punch-Out catalog, same process will have to be done.

Regards.

[laurent_burtaire]

Hello,

regarding Ariba catalog extraction, correct URL to use is https://s1-eu.ariba.com/Buyer/downloadJson/

In our case, s1-eu needs to be used as we are on the EU datacenter.

User name selection field does not have to be populated because in HTTP request, parameter expected by Ariba is “UserId” and not "USERNAME". So, I added UserId in catalog call structure.

Now, we are able to reach Ariba server, but we have a new problem being an Ariba internal server error.

Regards.

Laurent.

[laurent_burtaire]

Hello Ricardo,

thank you for your answer.

Regarding your first point for page size. I tried with 1500 instead of 20. I have the same authentication error sent back by APC.

Concerning SMICM, i already checked if i have error message: there is no problem.

Previously there was SSL error due to missing certificate: this point is solved.

I think the problem is on Ariba side, but i do not know which one.

We are waiting for answer from an Ariba consultant.

Regards.

Laurent.

[ricardo_cavedini]

Hello,

Which page size parameter are you using? For Ariba it should be 1500, as note 2380979 mentions.

Also, check the ICM trace for this extraction. Maybe there is a certificate issue when connecting to Ariba.

As these traces are really (really) long, you can follow these steps to obtain it:

- Go to transaction SMICM
- Navigate to "Goto" > "Trace Level" and set trace to level 2

- Navigate to "Goto" > "Trace File" and * the last line shown in trace file
- Reproduce the issue

- Go back to SMICM and check just the traces shown after the line marked previously

This trace should show an SSL error or something similar.

You will find more details about this kind of ICM errors in the SAP Note:

2244483 - Catalog errors in shopping cart after item was transferred

Regards,
Ricardo

[laurent_burtaire]

Hi again,

i checked in Ariba Procurement Content groups assigned to user i populated for field username: he has 'Catalog Extract'.

It seems not to be sufficient.

Regards.

Laurent.

[laurent_burtaire]

Hello,

regarding the point lv_body_string has no items detail, this was due to wrong parameters populated for HTTP/HTTPS extraction.

First, URL does not have to be the one coming from web Service call structure. Indeed, according to "Ariba Procurement Content Integration Guide", paragraph 'Extracting Catalog Data in JSON Format', correct URL is:

https://s1.ariba. com/Buyer/downloadJson/<ANID> where <ANID> is the Ariba Network ID value for Ariba Procurement Content.

Regarding user name and password, i tried with one "functionnal" user (having authorizations to connect to APC).

Now, data in lv_body_string iserror HTTP Status 401: This request requires HTTP authentication.

I think i do not have the correct user and password.

Regards.

Laurent.

[laurent_burtaire]

Hello,

other question: the multi-backend is not managed for HTTP/HTTPS catalog data extraction.

It is OK for Product Master as all material extracted come from table COMMPR01.

But for Punch-Out catalog, only one category GUID can be extracted.

Does anybody know if it is SAP limitation process or a bug?

Thank you.

Regards.

Laurent.

[laurent_burtaire]

Hi,

it seems it is necessary to make catalog data extraction for each backend concerned.

For this, you should have to use the catalog logsys field in the Web Service details.

Regards.

Laurent.