make sure all 3 backends hash the same

Author: nystrom

meta/src/meta/templates/parser.jl.template

@@ -228,9 +228,9 @@ end
 function relation_id_from_string(parser::ParserState, name::String)
     # Create RelationId from string and track mapping for debug info
     hash_bytes = sha256(name)
-    # Match the convention in backir-to-lqp.jl
-    id_low = reinterpret(UInt64, hash_bytes[1:8])[1]
-    id_high = reinterpret(UInt64, hash_bytes[9:16])[1]
+    # Use little-endian to match the Go and Python SDKs.
+    id_low = htol(reinterpret(UInt64, hash_bytes[1:8])[1])
+    id_high = htol(reinterpret(UInt64, hash_bytes[9:16])[1])
     relation_id = Proto.RelationId(id_low, id_high)
 
     # Store the mapping for the current fragment if we're inside one

sdks/go/src/hash_test.go

@@ -0,0 +1,16 @@
+package lqp
+
+import "testing"
+
+// TestRelationIdFromString verifies that relation_id_from_string produces
+// the same id across all SDKs (Julia, Python, Go).
+func TestRelationIdFromString(t *testing.T) {
+	p := NewParser([]Token{})
+	rid := p.relationIdFromString("my_relation")
+	if rid.IdLow != 0x5d33996702404f85 {
+		t.Errorf("id_low: got 0x%016x, want 0x5d33996702404f85", rid.IdLow)
+	}
+	if rid.IdHigh != 0x3b9af8e72af633f8 {
+		t.Errorf("id_high: got 0x%016x, want 0x3b9af8e72af633f8", rid.IdHigh)
+	}
+}

sdks/julia/LogicalQueryProtocol.jl/src/parser.jl

@@ -254,9 +254,9 @@ end
 function relation_id_from_string(parser::ParserState, name::String)
     # Create RelationId from string and track mapping for debug info
     hash_bytes = sha256(name)
-    # Match the convention in backir-to-lqp.jl
-    id_low = reinterpret(UInt64, hash_bytes[1:8])[1]
-    id_high = reinterpret(UInt64, hash_bytes[9:16])[1]
+    # Use little-endian to match the Go and Python SDKs.
+    id_low = htol(reinterpret(UInt64, hash_bytes[1:8])[1])
+    id_high = htol(reinterpret(UInt64, hash_bytes[9:16])[1])
     relation_id = Proto.RelationId(id_low, id_high)
 
     # Store the mapping for the current fragment if we're inside one

sdks/julia/LogicalQueryProtocol.jl/test/parser_tests.jl

@@ -112,6 +112,16 @@ end
     @test r.scale == 1
 end
 
+@testitem "Parser - relation_id_from_string" setup=[ParserSetup] begin
+    using LogicalQueryProtocol.Parser: ParserState, Token, relation_id_from_string
+
+    # All SDKs must produce the same id for the same string.
+    parser = ParserState(Token[])
+    rid = relation_id_from_string(parser, "my_relation")
+    @test rid.id_low == 0x5d33996702404f85
+    @test rid.id_high == 0x3b9af8e72af633f8
+end
+
 @testitem "Parser - Lexer tokenization" setup=[ParserSetup] begin
     lexer = Lexer("(transaction (epoch (writes) (reads)))")
     # Tokens: ( transaction ( epoch ( writes ) ( reads ) ) ) $

sdks/python/tests/test_parser.py

@@ -3,11 +3,19 @@
 import pytest
 from pytest_snapshot.plugin import Snapshot
 
-from lqp.gen.parser import parse
+from lqp.gen.parser import Parser, parse
 
 from .utils import BIN_SNAPSHOTS_DIR, get_lqp_input_files
 
 
+def test_relation_id_from_string():
+    """All SDKs must produce the same id for the same string."""
+    parser = Parser([])
+    rid = parser.relation_id_from_string("my_relation")
+    assert rid.id_low == 0x5D33996702404F85
+    assert rid.id_high == 0x3B9AF8E72AF633F8
+
+
 @pytest.mark.parametrize("input_file", get_lqp_input_files())
 def test_parse_lqp(snapshot: Snapshot, input_file):
     """Test that each input file can be parsed and matches its binary snapshot."""

tests/bin/cdc.bin

@@ -19,30 +19,36 @@
                :csv_skip 0})
             (columns
               (column "raw_data" [])
-              (column ["append" "METADATA$KEY"] 0xb7219c9b1f7d843d [UINT128])
-              (column ["delete" "METADATA$KEY"] 0xc45b4cf618eed32f [UINT128])
-              (column ["append" "user_id"] 0x7982cff88800e869 [INT])
-              (column ["append" "name"] 0x8cb6c4889f2f67c7 [STRING])
-              (column "id" 0x5364bf081a053f95 [INT])
-              (column ["append" "id"] 0x22d32663113e8a7d [INT])
-              (column ["delete" "id"] 0x337b9038b6bf2f6a [INT]))
+              (column
+                ["append" "METADATA$KEY"]
+                0x8dfeb4bd8f6d8c353d847d1f9b9c21b7
+                [UINT128])
+              (column
+                ["delete" "METADATA$KEY"]
+                0xa91fbfb76393ae562fd3ee18f64c5bc4
+                [UINT128])
+              (column ["append" "user_id"] 0xf68f290595d54fb669e80088f8cf8279 [INT])
+              (column ["append" "name"] 0x2756fdf898b5a25ac7672f9f88c4b68c [STRING])
+              (column "id" 0xc0c8b0135d8dee07953f051a08bf6453 [INT])
+              (column ["append" "id"] 0xd21b8166f6243157d8a3e116326d322 [INT])
+              (column ["delete" "id"] 0xe870ce327bf8f16e6a2fbfb638907b33 [INT]))
             (asof "2025-06-01T00:00:00Z")))))
     (reads
-      (output :mk_ins 0xb7219c9b1f7d843d)
-      (output :mk_del 0xc45b4cf618eed32f)
-      (output :uid_ins 0x7982cff88800e869)
-      (output :name_ins 0x8cb6c4889f2f67c7)
-      (output :entity_id 0x5364bf081a053f95)
-      (output :eid_ins 0x22d32663113e8a7d)
-      (output :eid_del 0x337b9038b6bf2f6a))))
+      (output :mk_ins 0x8dfeb4bd8f6d8c353d847d1f9b9c21b7)
+      (output :mk_del 0xa91fbfb76393ae562fd3ee18f64c5bc4)
+      (output :uid_ins 0xf68f290595d54fb669e80088f8cf8279)
+      (output :name_ins 0x2756fdf898b5a25ac7672f9f88c4b68c)
+      (output :entity_id 0xc0c8b0135d8dee07953f051a08bf6453)
+      (output :eid_ins 0xd21b8166f6243157d8a3e116326d322)
+      (output :eid_del 0xe870ce327bf8f16e6a2fbfb638907b33))))
 
 ;; Debug information
 ;; -----------------------
 ;; Original names
-;; 	 ID `0x337b9038b6bf2f6a` -> `eid_del`
-;; 	 ID `0x22d32663113e8a7d` -> `eid_ins`
-;; 	 ID `0x5364bf081a053f95` -> `entity_id`
-;; 	 ID `0xc45b4cf618eed32f` -> `mk_del`
-;; 	 ID `0xb7219c9b1f7d843d` -> `mk_ins`
-;; 	 ID `0x8cb6c4889f2f67c7` -> `name_ins`
-;; 	 ID `0x7982cff88800e869` -> `uid_ins`
+;; 	 ID `0xe870ce327bf8f16e6a2fbfb638907b33` -> `eid_del`
+;; 	 ID `0xd21b8166f6243157d8a3e116326d322` -> `eid_ins`
+;; 	 ID `0xc0c8b0135d8dee07953f051a08bf6453` -> `entity_id`
+;; 	 ID `0xa91fbfb76393ae562fd3ee18f64c5bc4` -> `mk_del`
+;; 	 ID `0x8dfeb4bd8f6d8c353d847d1f9b9c21b7` -> `mk_ins`
+;; 	 ID `0x2756fdf898b5a25ac7672f9f88c4b68c` -> `name_ins`
+;; 	 ID `0xf68f290595d54fb669e80088f8cf8279` -> `uid_ins`