DOCDB - putting a square peg in a round hole

Part II – exotic kind codes and number suffixes: the details

In the world of IP, patent information leads to patent knowledge and everything we know about the state of the art.

Jurisdictions that exchange data with the EPO have great interest in loading their data to DOCDB and having it presented correctly. It is also essential that EPO patent examiners have all this information at hand to carry out search and examination to the high standards the EPO is known for. Our data experts are of course very much aware of this and work constantly to reduce data gaps and keep information up to date.

This second part on exotic kind codes looks in more detail at the codes and suffixes you may come across in application or priority numbers and also gives you some historical background.

Kind codes D and Q

These are internal codes used to identify applications with incomplete bibliographic data. They can be assigned for different reasons – for records that do not follow WIPO publication standards, for collections delivered in a different format, or where information such as dates or kind codes is missing. For example, EPO databases contain very early publications available only in PDF or image format, but without bibliographic data. In the past, and for EPO internal business needs, these publications were recorded using only the publication number. As a workaround, the application number was made up of the publication number and the kind code D. In some cases, this same application number was also used in place of a missing date (this is not to be mistaken for an error in the bibliographic data – see screenshot below).

Example from Espacenet image 1
Figure 1. Example of application number and date derived from publication number. (click on the image to enlarge)

Note: D and Q are purely used internally to help identify necessary corrective actions:

  • Application kind-code D issues can only be resolved through an automated back file correction exercise.
  • Application kind-code Q issues require intellectual effort and are fixed manually one by one.

Application codes K, L, M, N and O

Some countries, e.g., MC, PH, RU and SU, have used and in some cases still use identical application numbers for different publications. To make them unique, these codes are added to each application number in alphabetical order starting with K.

Example from Global Patent Index image 2
Figure 2. Application number MC1374 with internal code K.

Example from Global Patent Index image 3
Figure 3. Application number MC1374 with internal code L.

Old French applications with kind-codes E, F and M

These codes are applied to French applications between 1900 and 1960s that were loaded with incorrect kind codes and without dates in XML. Some shared identical application numbers with other publications, making it impossible to distinguish between them. In order to make such publications which exist only in PDF format visible in tools like Espacenet, we needed to create a bibliographic record first. For this purpose, we extracted only the publication number from the document and appended one of the above-mentioned kind codes. This newly created bibliographic record – even if it consists only of a single number– makes it possible for the user to access this document in Espacenet.

Espacenet with kind code image 4
Figure 4. Publication number FR99 with suffix E.

Espacenet with kind code image 5
Figure 5. Publication number FR99 with suffix F.

Numbers suffixed with D or T

Application numbers suffixed with D and priority numbers suffixed with T are "dummy" entries that are eventually replaced with genuine identifiers as soon as the bibliographic data is made available to the EPO. This only happens if the jurisdiction in question provides a new backfile. When loaded, this then triggers a renumbering process in which dummy entries are replaced with genuine identifiers. This explains why the number of records with dummy suffixes in EPO databases has decreased over the years.

Numbers suffixed with T

Very early publications provided in paper form were recorded using publication numbers only. Wherever possible, EPO examiners grouped these publications into patent families after studying the technical content and creating a dummy priority from one of the family publication numbers and adding the T suffix.

Example from Global Patent Index image 6
Figure 6. Example from Global Patent Index of a T priority number suffix.

Numbers suffixed with X

These are priorities claimed in publications from the early 20th century and are often found in US documents. Back then, priority claims came without a number and could only be identified using the country and date. As DOCDB requires a priority number for each record to create a patent family, the EPO generated a priority number consisting of the publication number with suffix X.

For users of DOCDB in XML, it's important to keep in mind that the following special format applies to numbers that are suffixed with D, T or X:

  • country code
  • D or T or X
    • number
    • kind code when kind code not = 'A'

As you can see, EPO experts try hard to make as much patent information available to you as soon as possible. If you ever need help with kind codes, or with any other data-related topic, email our team of experts at

Further information

Keywords : DOCDB, document kind codes, number suffixes

Quick Navigation