Why is the section letter not in the numeric code?

The section letter is not in the numeric code because the regulation defines the code as the division, group and class numbers, with the section as a separate level above them. The Eurostat manual states that the code for the section level is not integrated in the NACE code that identifies the division, the group and the class describing a specific activity. The section must therefore be found by looking up the division in the structure table.

How does the regulation describe the four levels?

The separation is a consequence of the hierarchy design. Each lower category belongs to exactly one category above, so the division number already determines the section, once the structure table is consulted. Adding the letter to the code would duplicate information that the structure already contains, and it would make the code longer without adding a distinction. The design therefore keeps the code numeric, and keeps the section as a lookup. The trade off is that a numeric code on its own does not name its section, and a reader must do the lookup rather than reading the section from the code.

Identifier shown	Level	In the numeric code
One letter, such as C	Section	No, found by lookup
Two digits, such as 20	Division	Yes, first two digits
Three digits, such as 20.5	Group	Yes, first three digits
Four digits, such as 20.51	Class	Yes, all four digits

What happens when sections shift letters?

When sections shift letters, the numeric codes can stay the same while the section letter changes. That is what happened after the split of the former Section J. The manual Table 4 shows the comparison: the former section K for financial and insurance activities became section L, the former section L for real estate activities became section M, and each following section shifted one letter, through to the former section U becoming section V. Many division numbers inside those sections stayed the same, because the shift was at section level, not a renumbering of divisions. A reader who records a section letter from the earlier version beside a numeric code from the later version, or the reverse, will therefore record a wrong section for that code.

The shift also shows why datasets must state the version beside the code and beside any section letter shown. A table that shows a section letter and a numeric code together is showing a presentation, derived from the structure table for a stated version. A table that shows only numeric codes is showing the code as defined, and the section must be derived from the structure table for the version stated for the table. A code with a section letter from a different version attached to it is a mixed version presentation, and it must be corrected by re deriving the section from the correct version structure.

How are codes presented with sections in practice?

In practice, codes are often presented with the section letter beside them, in a separate column, or as a prefix in a display label. The Central Statistics Office in Ireland, for example, presents sector groupings by section letter in its background notes, such as industry as sections B to F, while the underlying codes remain numeric. That presentation is legitimate, provided the version is stated and the letter is derived from the structure table for that version. The presentation does not change the code. A reader who copies a displayed label that includes a letter must check whether the letter is part of the stored code or a display addition, before using the copied value as a code in a search or a join with another dataset.

The same care applies to the dot in the code. The manual states that NACE places a dot between the first two digits and the last two digits, to distinguish the coding from that used in the International Standard Industrial Classification of All Economic Activities. A stored code without a dot, such as a four digit string, is the same code in a different storage form, and the level boundaries must be restored by position: first two digits are the division, third digit is the group digit, fourth digit is the class digit. The section letter is not restored by position, because it is not in the stored code in any form. It is restored only by lookup.

What does the separation not mean?

The separation does not mean that the section is unimportant, or that a class can be read without knowing its section. The section note states the broad scope and the main boundaries, and the class note must be read inside that scope. A class in manufacturing and a class with a similar heading in another section are different categories, even if their headings share words, because their parent sections, divisions and groups differ, and their notes state different inclusions and exclusions. The section lookup is therefore a necessary step in reading a code, even though the letter is not in the code. Skipping the lookup risks reading the class note outside its section context.

Look up the division in the structure table to find the section, read the section note and the class note together, and confirm any entry with the national statistical institute or the business register that holds the record.