How does the Section J split appear in a correspondence table?

The Section J split appears in a correspondence table as a set of class level moves between NACE Rev. 2 and NACE Rev. 2.1, recorded under the four correspondence types the manual defines. Paragraph 207 of the introductory guidelines defines the types and gives their counts for the whole classification. A correspondence table records moves. It does not pick a new class for a specific unit, and it does not guarantee a one to one result.

What are the four correspondence types?

Paragraph 207 states that the correspondences between NACE Rev. 2 and NACE Rev. 2.1 are of four types, in correspondence table version 1.05 as cited in the manual. A one to one correspondence covers 354 Rev. 2 classes which correspond exactly to one class in Rev. 2.1 and vice versa. A many to one correspondence covers 25 Rev. 2 classes merged into one class in Rev. 2.1. A one to many correspondence covers 11 Rev. 2 classes split into two or more classes in Rev. 2.1. A many to many correspondence covers all other Rev. 2 classes, 225 in all, for which the joint scope of two or more Rev. 2 categories is identical to the joint scope of two or more Rev. 2.1 categories.

The counts add to the class totals of the two versions in a way that illustrates the structure of the change. The 354 one to one classes are the unchanged core. The 25 many to one classes are the mergers. The 11 one to many classes are the clean splits. The 225 many to many classes are the remainder, where content has been rearranged across several categories at once, so that no single old class maps cleanly to a single new class and no single new class maps cleanly back. The Section J split, which rearranged divisions, groups and classes together, belongs in large part to the rearranged remainder, and a reader should expect many to many patterns in that area rather than clean one to one results.

How is a split shown in the table?

A split is shown as one source row pointing to two or more target rows, in the one to many case, or as part of a block of source rows pointing to a block of target rows, in the many to many case. The table does not add a column that says which target row applies to a specific unit. It records the scopes. The explanatory note for each target class states what the class includes and excludes, and the note is the document that allows a reader to see which target scope contains the activity in question, in general terms. The correspondence table and the explanatory notes are therefore companion documents, and the manual presents them together: paragraph 38 states that Eurostat published explanatory notes as well as correspondence tables, and paragraph 206 states that complete detailed correspondences in both directions are available online.

The version of the table matters. Paragraph 207 cites correspondence table version 1.05. A reader who uses a different version, or a national variant table, must state which version is being used, because the recorded moves can differ between versions. The Eurostat correspondence tables page, listed in the sources, is the entry point for the current tables. National statistical institutes publish their own correspondence arrangements for national variants, and those arrangements must be checked with the institute concerned, not assumed from the Eurostat table alone.

What must a reader not infer from a table row?

| Correspondence type | What the table shows | What the table does not show |

|---|---|---|

| One to one | One old class, one new class, exact correspondence | Nothing further is needed at scope level, but the note must still be read |

| Many to one | Several old classes merged into one new class | Which old class a specific unit came from, as a current entry |

| One to many | One old class split into two or more new classes | Which new class applies to a specific unit |

| Many to many | Joint scope of several old categories equals joint scope of several new categories | Any single mapping for a specific unit |

The table follows paragraph 207. The right hand column is the limit that governs the whole cluster: a correspondence table records scopes, and the entry for a specific unit is confirmed elsewhere. The guardrail on this page states the same limit for the reader directly. A table row quoted without its type, and without the note for the target class, is an incomplete citation, and it must not be presented as a determination.

What does the split not decide?

The split changes the structure of the classification. A code legal effect is unchanged by this, because beyond statistical use a code has no legal effect. Eurostat and the EUR-Lex summary both confine the legislation to statistical purposes. When an old Section J code appears in a dataset, the classification version the dataset uses and the correspondence table account of the move are the two questions that come first. The letter alone answers neither question, and this guide answers neither for a specific unit.

Where is a moved code confirmed?

Confirmation here means three checks in order. First, the section 4 summary in the manual for the structural change. Second, the explanatory notes for the categories at each end of the move. Third, the institute or register for the entry held for a specific unit. This guide gives the first check and points to the other two.

Read the table type first, read the note for each target class second, and confirm the entry for a specific unit with the institute or register that holds it.