Record matching rules are used when importing MARC records into Koha. Koha includes some Record Matching Rules, such as ISBN, ISSN, 999$c (Koha's Unique Identifier), and 035 (OCLC number), by default. During an import, a staff member can select which matching rule should be applied for an import.
Match Points
Setting up the match points is the first step (and potentially the only step) in creating a matching rule. A match point is a requirement that must be met in order for an incoming record to match an existing MARC record in your catalog. You can have multiple match points on an import rule, each with its own score. An incoming record will be compared against your existing records (‘one record at a time’) and given a score for each match point. When the total score of the match points matches or exceeds the threshold given for the matching rule, Koha assumes a good match and imports/overlays according your specifications in the import process.
Match points can be used for both Bibliographic and Authority Imports.

Rules that match on 020$a will look at all instances of 020$a in the incoming bib, not just the first. Further, if the system preference AggressiveMatchOnISBN is set to "When matching on ISBN with the record import tool, [do] attempt to match aggressively by trying all variations of the ISBNs in the imported record as a phrase in the ISBN fields of already cataloged records", Koha will remove extraneous text from the incoming 020$a and convert ISBN-10 to ISBN-13 (and vice versa) as needed.
Match Check
While match points work on the search index, match checks work directly on the data and can be used as the matching criteria instead of the match points or in addition to them to confirm true matches. With match checks, a library can use one MARC field in the incoming record and match it on another MARC record in the database.
Here is the match checks we used in the example, first the match point was set to ISBN, and then we want Koha to then look at the Title for a fuller match: