Using appln_auth = 'MX' is about patents filed at the Mexican patent office. Using the person country is about applicants (or inventors) that have their place of business in Mexico. One can of course also combine the 2 conditions. "Patents about coffee" might need some fine tuning, but IPC...

@ i.dijkstra.8 There are no functional restrictions exept for the download volumes and the limitation to 1 month free use. Your first query works fine. (no errors) The second one had a small error due to inconsistent use of the tls201_appln alias (t1). From a concept point of view, if you want to me...

@khanhhienbi, you can count distinct docdb families, but then you will loose the granularity of knowing where the patents were filed. It is 2 different things: number of families is a proxy for "number of inventions". Family size is an indication of the geographical scope of where the inve...

For details explanation about the docdb_family_size, kindly see that PATSTAT data catalog. The "logics" in the query are not 100% correct, because a patent family can have patents filed in different years; and the family members will be filed (mostly) in different countries. So having the ...

Espacenet has a limitation in the download of 500 families. You can use the family number in PATSTAT via the docdb_fam_id in tls201_appln. If you want to have more then 500 then you will need to chunk it up in ESPACENET. I mostly use the "Earliest publication date (family)" in the filter t...

Using WHERE LEFT(cpc_class_symbol, 3) = 'Y02' OR LEFT(cpc_class_symbol, 4) IN ('Y02A', 'Y02B', 'Y02C', 'Y02D', 'Y02E', 'Y02P', 'Y02T', 'Y02W') is overlapping; the LEFT(cpc_class_symbol, 4) will not add anything as you allready take all "Y02"s. But the reason for the limited number of resul...

No, your query does not take into account any triadic concept. The concept "where does innovation takes place" often uses the country where a priority (or first filing) was filed. Some researchers would use the country of the inventors to answer "where does innovation take place"...

Hello, the change was not so unexpected :) It happened for the 2021 Spring release. The [cited_npl_publn_id] attribute should indeed be 32 characters long, NOT integer. The data catalog indicates this also in the Modification History for the [cited_npl_publn_id] but also in the "(7) History of ...

You could use something like this: SELECT DISTINCT a.appln_filing_year, a.appln_id, a.appln_auth, a.appln_nr, a.appln_kind, a.appln_filing_date, STRING_AGG (CPC.cpc_class_symbol, '; ') CPC_GROUP FROM tls201_appln a JOIN tls225_docdb_fam_cpc CPC ON a.docdb_family_id = CPC.docdb_family_id WHERE left(c...

The service fee (150 euro) for PATSTAT Global (or any of the data sets on the download platform) is abolished for 2024 subscriptions. But you still need to register to get access to the platform. (user_id, password, etc...) PATSTAT is produced 2 times per year. We are currently working on the Spring...