19 / 20 Reading NACE codes in registers, datasets and listings
How are code formats, punctuation and spacing read?
Published Updated
Statistical classification only
Statistical classification only. This guide does not determine any reader's code. Confirm a code with the national statistical institute or the business register that holds the record. This is general information, not advice on any reader's registration.
Code formats use letters and digits in fixed positions, and punctuation and spacing are presentation, not part of the category. A section is one letter. A division is two digits. A group is three digits. A class is four digits. Those formats are stated in the EUR-Lex summary, in the Eurostat NACE background article, and in the Central Statistics Office account of the hierarchy.
A code should be copied exactly as shown in the source, including any dot, space, or national digit or letter, and it should not be reformatted without a note, because reformatting can hide a level, a version, or a national detail.
What do dots and spaces mean in a code?
Dots and spaces in a code are separators between levels, added for readability, and they do not change the category. The Dutch SBI excerpt published by Statistics Netherlands shows the pattern: a group shown as 01.1, a class shown as 01.11, and a fifth level shown as 01.11.0, each with its title. The dots separate the division, group, class and fifth level parts, and the code without dots, 01110, would be the same category in a system that stores codes without separators. A reader who searches for a code should try the format shown in the source, and, where the search fails, the format without separators, and should record which format was used. A reader who compares two codes should compare the digits and letters, not the punctuation, and should state the format used in the comparison, so that a difference in punctuation is not mistaken for a difference in category. Punctuation is presentation. The digits and letters are the code.
What do national digits and letters mean?
National digits and letters are detail added below the European class, for national use, and they must be read as national detail, not as European categories. The Dutch SBI has a fifth level, with a maximum of five digits, and the fifth level is specific to the Netherlands, as Statistics Netherlands states. The United Kingdom SIC adds a five digit sub class level for certain classes, as the Office for National Statistics states, below the four digit class level shared with NACE. The French APE code is shown with an activity label in the Sirene record, and the code format in the French record includes the national detail for the French version, as Service Public describes in its account of the APE code and the Sirene tool. A national code should therefore be read in two parts: the shared part, to the fourth digit, which is comparable at European level, and the national part, which is comparable only within the national version. A confirmation should state how many characters the code as held contains, and which part is national, so that a comparison does not treat national detail as a European difference.
How are section letters read after the J and K split?
Section letters are read in the lettering of the version in use, not in the lettering of an earlier version. NACE Rev. 2.1 split the former Section J, information and communication, into Section J, publishing, broadcasting and content production and distribution, and Section K, telecommunication, computer programming, consulting, computing infrastructure and other information service activities, as the Central Statistics Office states, and the sections that followed shifted one letter as a result. A section letter in a NACE Rev. 2.1 record therefore does not mean the same as the same letter in a NACE Rev. 2 record, where the lettering differs, and a reader must check the version before reading the letter. A table that shows section letters without stating the version is ambiguous, and it should be completed with the version from the dataset metadata or the report, before any comparison is made. The letter is part of the code. The version is part of the meaning of the letter.
| Format element | Meaning |
|---|---|
| One letter | Section, broadest grouping |
| Two digits | Division, inside a section |
| Three digits | Group, inside a division |
| Four digits | Class, most detailed European level |
| Fifth digit or national letter | National detail, below the class, version specific |
| Dot or space | Separator between levels, presentation only |
The table shows the format elements in order, from the broadest to the most detailed, with the national element and the presentation element separated at the end. A code read with that table will not confuse a separator with a level, or national detail with a European category.
How is a code copied without error?
A code is copied without error by copying the characters, not by retyping from memory, and by checking the copy against the source before quoting it. A single digit error changes the category, and a single letter error changes the section, and neither error is visible in the title alone, because the title is not always shown with the code. A reader should copy the code from the source, paste it where the system allows, and, where typing is necessary, read it back digit by digit against the source. A code that is quoted from a listing, a PDF, or a dataset should be checked in the register or the code list for the version stated, before it is reused, because a copying error in the listing or the PDF will be repeated in every quotation that follows it. The check is cheap, and the error it prevents is expensive, because a wrong code quoted repeatedly acquires the appearance of a verified code.
What does a format not prove?
A format does not prove the version, the meaning, or the entry. A four digit code in the correct format can belong to NACE Rev. 2 or NACE Rev. 2.1, and it can be a class in one and a different class in the other, where the category moved. A five digit code can be a Dutch SBI code, a United Kingdom SIC sub class, or a code in another national version, and the format alone does not distinguish them. The format is the shape of the code. The version, the note, and the holder are the meaning and the record, and they must be stated alongside the format in every quotation. A code in the correct format, without a version, is a well formed string, not a checked category.
Copy the code exactly as shown, state the version and the national detail, and confirm the entry with the register that holds the record.