EPO - Discussion forums

Posted: **Mon Sep 21, 2020 11:31 pm**

Dear PATSTAT support,

I am using the one-month trial of PATSTAT Online and I was experimenting on its processing capabilities for aggregate statistical analysis in lights of a future purchase. However, whenever I tried to process more than 2 million observations and select the top elements, I receive this error message as a response:

Error 1014: service unavailable, please try again later
Error 500 Internal Server Error
f@https://data.epo.org/expert-services/do ... .js:80:437
f@https://data.epo.org/expert-services/do ... js:106:199
e@https://data.epo.org/expert-services/do ... js:109:229

Is this a temporary issue? I tried with different queries but the result is the same.

Thank you

Al

Posted: **Tue Sep 22, 2020 8:07 am**

Dear Al,
PATSTAT Online has indeed a cut-off when the computational cost of a query becomes too high.
Although the platform is MS SQL based, some functions or features might not work 100% because the parser pre-treats it before sending it to the server.
Can you post the query so we can have a look at it ?
Geert BOEDT

Posted: **Tue Sep 22, 2020 10:10 am**

Here it is. I am trying to list the major applicants in a technical field for the year 2019

Code: Select all

SELECT DISTINCT pers.psn_name, p.appln_id, p.publn_auth, p.publn_nr, p.publn_kind, p.publn_date
FROM tls211_pat_publn p
JOIN tls209_appln_ipc i ON p.appln_id = i.appln_id
JOIN tls207_pers_appln pa ON p.appln_id = pa.appln_id
JOIN tls206_person pers ON pa.person_id = pers.person_id 
WHERE i.ipc_class_symbol LIKE 'G%' -- Change IPC symbol here
AND YEAR(p.publn_date) = 2019
AND pa.applt_seq_nr > 0 
ORDER BY pers.psn_name, p.appln_id

Posted: **Tue Sep 22, 2020 2:03 pm**

Hello Al,
That query should (and does) run without error message and gives you a list of 1.590 million rows.
I assume it must have been a temporary hick-up - if this gave an error-.

But this query creates for sure too much records to download in a result list or to extract in a separate database. The conditions on the query are very generous.

If you want to create a ranking list: you could use something like this - or alternative- you count the number of Docdb families instead of applications in order to avoid getting noise from patents being filed in many countries.

Code: Select all

SELECT DISTINCT pers.psn_name,  count(distinct(p.appln_id)) total
FROM tls211_pat_publn p
JOIN tls209_appln_ipc i ON p.appln_id = i.appln_id
JOIN tls207_pers_appln pa ON p.appln_id = pa.appln_id
JOIN tls206_person pers ON pa.person_id = pers.person_id
WHERE i.ipc_class_symbol LIKE 'G%' 
AND YEAR(p.publn_date) = 2019
AND pa.applt_seq_nr > 0 and invt_seq_nr = 0
group by pers.psn_name
ORDER BY total desc

Geert BOEDT

EPO - Discussion forums

ERROR 1014

ERROR 1014

Re: ERROR 1014

Re: ERROR 1014

Re: ERROR 1014