Cypher & Neo4J compatibility : missing functions

[ExtReMLapin]

Hello,

The following functions are missing :

• coll.distinct
• coll.flatten
• coll.indexof
• coll.insert
• coll.max
• coll.min
• coll.remove
• coll.sort
• elementid
• tobooleanlist
• tofloatlist
• tointegerlist
• tostringlist
• file
• linenumber
• exists
• btrim
• lower (alias)
• upper (alias)
• normalize
• vector
• vector_dimension_count
• vector_distance
• vector.distance.euclidean
• vector.norm

whole adoc documentation for AI bots that strugles at webfetch

list.txt
load-csv.txt
mathematical-logarithmic.txt
mathematical-numeric.txt
mathematical-trigonometric.txt
predicate.txt
scalar.txt
spatial.txt
string.txt
user-defined.txt
vector.txt
aggregating.txt
database.txt
graph.txt
index.txt

[ExtReMLapin]

Linked pr that test those functions #3428

[lvca]

Implementation Status

23 out of 25 functions have been implemented. Here's the updated checklist:

• coll.distinct
• coll.flatten
• coll.indexof
• coll.insert
• coll.max
• coll.min
• coll.remove
• coll.sort
• elementid (was registered but not wired — now fully working)
• tobooleanlist
• tofloatlist
• tointegerlist
• tostringlist
• file — LOAD CSV-specific context function, requires LOAD CSV infrastructure
• linenumber — LOAD CSV-specific context function, requires LOAD CSV infrastructure
• exists
• btrim (alias for trim)
• lower (alias for toLower)
• upper (alias for toUpper)
• normalize (supports NFC, NFD, NFKC, NFKD)
• vector (alias for vector_create)
• vector_dimension_count
• vector_distance (supports EUCLIDEAN, COSINE, MANHATTAN metrics)
• vector.distance.euclidean
• vector.norm

All coll.* functions are also accessible via apoc.coll.* prefix for Neo4j/APOC compatibility.

The remaining file() and linenumber() are tied to LOAD CSV execution context and will be tracked separately if needed.

[lvca]

I was able to implement all the remaining. Now we have a jungle of OpenCypher and SQL functions that overlaps. I'm gonna refactor soon (no changes from user perspective), and maybe unify the functions in one big package function outside the SQL and OpenCypher with subpackage per type, like "math", "vector", etc. This is partially done, we need more cleaning, spring is coming!

[ExtReMLapin]

Running the tests again, thanks !

[lvca]

CSV related are more complicated, I'll move them to another issue