Initialize sqlparse lexer and tweak order of setting compilation fields

.changes/unreleased/Fixes-20230726-104448.yaml

@@ -0,0 +1,6 @@
+kind: Fixes
+body: Improve handling of CTE injection with ephemeral models
+time: 2023-07-26T10:44:48.888451-04:00
+custom:
+  Author: gshank
+  Issue: "8213"

core/dbt/compilation.py

@@ -4,7 +4,6 @@
 import networkx as nx  # type: ignore
 import os
 import pickle
-import sqlparse
 
 from collections import defaultdict
 from typing import List, Dict, Any, Tuple, Optional
@@ -36,6 +35,7 @@
 from dbt.events.format import pluralize
 import dbt.tracking
 import dbt.task.list as list_task
+import sqlparse
 
 graph_file_name = "graph.gpickle"
 
@@ -378,16 +378,16 @@ def _recursively_prepend_ctes(
 
             _add_prepended_cte(prepended_ctes, InjectedCTE(id=cte.id, sql=sql))
 
-        injected_sql = inject_ctes_into_sql(
-            model.compiled_code,
-            prepended_ctes,
-        )
         # Check again before updating for multi-threading
         if not model.extra_ctes_injected:
+            injected_sql = inject_ctes_into_sql(
+                model.compiled_code,
+                prepended_ctes,
+            )
+            model.extra_ctes_injected = True
             model._pre_injected_sql = model.compiled_code
             model.compiled_code = injected_sql
             model.extra_ctes = prepended_ctes
-            model.extra_ctes_injected = True
 
         # if model.extra_ctes is not set to prepended ctes, something went wrong
         return model, model.extra_ctes
@@ -563,6 +563,12 @@ def inject_ctes_into_sql(sql: str, ctes: List[InjectedCTE]) -> str:
     if len(ctes) == 0:
         return sql
 
+    # Make sure Lexer for sqlparse 0.4.4 is initialized
+    from sqlparse.lexer import Lexer  # type: ignore
+
+    if hasattr(Lexer, "get_default_instance"):
+        Lexer.get_default_instance()
+
     parsed_stmts = sqlparse.parse(sql)
     parsed = parsed_stmts[0]