sqlframe 3.4.0__py3-none-any.whl → 3.4.1__py3-none-any.whl

sqlframe/_version.py

@@ -12,5 +12,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '3.4.0'
-__version_tuple__ = version_tuple = (3, 4, 0)
+__version__ = version = '3.4.1'
+__version_tuple__ = version_tuple = (3, 4, 1)

sqlframe/base/dataframe.py

@@ -15,7 +15,7 @@ from dataclasses import dataclass
 
 import sqlglot
 from prettytable import PrettyTable
-from sqlglot import Dialect
+from sqlglot import Dialect, maybe_parse
 from sqlglot import expressions as exp
 from sqlglot import lineage as sqlglot_lineage
 from sqlglot.helper import ensure_list, flatten, object_to_dict, seq_get
@@ -460,16 +460,40 @@ class _BaseDataFrame(t.Generic[SESSION, WRITER, NA, STAT, GROUP_DATA]):
         df.expression.ctes[-1].set("cache_storage_level", storage_level)
         return df
 
-    @classmethod
-    def _add_ctes_to_expression(cls, expression: exp.Select, ctes: t.List[exp.CTE]) -> exp.Select:
+    def _add_ctes_to_expression(self, expression: exp.Select, ctes: t.List[exp.CTE]) -> exp.Select:
         expression = expression.copy()
         with_expression = expression.args.get("with")
         if with_expression:
             existing_ctes = with_expression.expressions
-            existsing_cte_names = {x.alias_or_name for x in existing_ctes}
+            existing_cte_counts = {x.alias_or_name: 0 for x in existing_ctes}
+            replaced_cte_names = {}  # type: ignore
             for cte in ctes:
-                if cte.alias_or_name not in existsing_cte_names:
-                    existing_ctes.append(cte)
+                if replaced_cte_names:
+                    cte = cte.transform(replace_id_value, replaced_cte_names)  # type: ignore
+                if cte.alias_or_name in existing_cte_counts:
+                    existing_cte_counts[cte.alias_or_name] += 10
+                    cte.set(
+                        "this",
+                        cte.this.where(
+                            exp.EQ(
+                                this=exp.Literal.number(existing_cte_counts[cte.alias_or_name]),
+                                expression=exp.Literal.number(
+                                    existing_cte_counts[cte.alias_or_name]
+                                ),
+                            )
+                        ),
+                    )
+                    new_cte_alias = self._create_hash_from_expression(cte.this)
+                    replaced_cte_names[cte.args["alias"].this] = maybe_parse(
+                        new_cte_alias, dialect=self.session.input_dialect, into=exp.Identifier
+                    )
+                    cte.set(
+                        "alias",
+                        maybe_parse(
+                            new_cte_alias, dialect=self.session.input_dialect, into=exp.TableAlias
+                        ),
+                    )
+                existing_ctes.append(cte)
         else:
             existing_ctes = ctes
         expression.set("with", exp.With(expressions=existing_ctes))
@@ -843,11 +867,11 @@ class _BaseDataFrame(t.Generic[SESSION, WRITER, NA, STAT, GROUP_DATA]):
             logger.warning("Got no value for on. This appears to change the join to a cross join.")
             how = "cross"
         other_df = other_df._convert_leaf_to_cte()
+        join_expression = self._add_ctes_to_expression(self.expression, other_df.expression.ctes)
         # We will determine actual "join on" expression later so we don't provide it at first
-        join_expression = self.expression.join(
-            other_df.latest_cte_name, join_type=how.replace("_", " ")
+        join_expression = join_expression.join(
+            join_expression.ctes[-1].alias, join_type=how.replace("_", " ")
         )
-        join_expression = self._add_ctes_to_expression(join_expression, other_df.expression.ctes)
         self_columns = self._get_outer_select_columns(join_expression)
         other_columns = self._get_outer_select_columns(other_df.expression)
         join_columns = self._ensure_and_normalize_cols(on)

{sqlframe-3.4.0.dist-info → sqlframe-3.4.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: sqlframe
-Version: 3.4.0
+Version: 3.4.1
 Summary: Turning PySpark Into a Universal DataFrame API
 Home-page: https://github.com/eakmanrq/sqlframe
 Author: Ryan Eakman

{sqlframe-3.4.0.dist-info → sqlframe-3.4.1.dist-info}/RECORD

@@ -1,10 +1,10 @@
 sqlframe/__init__.py,sha256=E3qCJ4PSEgKz6Hg3ves6LWt3JrQOV8c9HVLSIUOzKNc,3106
-sqlframe/_version.py,sha256=YWGqQYvejjlymmjzg4jncyBgDC760jlRmyon_Rd-2uQ,411
+sqlframe/_version.py,sha256=LQZJn_TZdGacASGuhVlLDchuihNUob4wWhR-2F34rXE,411
 sqlframe/base/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 sqlframe/base/_typing.py,sha256=b2clI5HI1zEZKB_3Msx3FeAJQyft44ubUifJwQRVXyQ,1298
 sqlframe/base/catalog.py,sha256=SzFQalTWdhWzxUY-4ut1f9TfOECp_JmJEgNPfrRKCe0,38457
 sqlframe/base/column.py,sha256=C2xj6OHMsJbEgjbI-m5HuIvqHYt2DbbUtCjssKpplNk,17748
-sqlframe/base/dataframe.py,sha256=uTM46auSPniww-ZMtI5s0hrrCG1nOICMEhdZZ8E4cu8,71636
+sqlframe/base/dataframe.py,sha256=coeUwntwYbT1g6YKVwk3ZfWMfJqAzd1ECYabBSsNsV0,72892
 sqlframe/base/decorators.py,sha256=Jy4bf8MhZ-AJ6CWTj59bBJRqamtLbPC0USUMFrY6g0w,449
 sqlframe/base/exceptions.py,sha256=9Uwvqn2eAkDpqm4BrRgbL61qM-GMCbJEMAW8otxO46s,370
 sqlframe/base/function_alternatives.py,sha256=IxNBqplehkAEkpzA625Dif-9Xyi4Hrho81A9U262rV0,50714
@@ -107,8 +107,8 @@ sqlframe/standalone/udf.py,sha256=azmgtUjHNIPs0WMVNId05SHwiYn41MKVBhKXsQJ5dmY,27
 sqlframe/standalone/window.py,sha256=6GKPzuxeSapJakBaKBeT9VpED1ACdjggDv9JRILDyV0,35
 sqlframe/testing/__init__.py,sha256=VVCosQhitU74A3NnE52O4mNtGZONapuEXcc20QmSlnQ,132
 sqlframe/testing/utils.py,sha256=9DDYVuocO7tygee3RaajuJNZ24sJwf_LY556kKg7kTw,13011
-sqlframe-3.4.0.dist-info/LICENSE,sha256=VZu79YgW780qxaFJMr0t5ZgbOYEh04xWoxaWOaqIGWk,1068
-sqlframe-3.4.0.dist-info/METADATA,sha256=IDBAwEcCrJfWwGuAaa8kfV3Jxlte38uOYO87zcFWJfY,8639
-sqlframe-3.4.0.dist-info/WHEEL,sha256=G16H4A3IeoQmnOrYV4ueZGKSjhipXx8zc8nu9FGlvMA,92
-sqlframe-3.4.0.dist-info/top_level.txt,sha256=T0_RpoygaZSF6heeWwIDQgaP0varUdSK1pzjeJZRjM8,9
-sqlframe-3.4.0.dist-info/RECORD,,
+sqlframe-3.4.1.dist-info/LICENSE,sha256=VZu79YgW780qxaFJMr0t5ZgbOYEh04xWoxaWOaqIGWk,1068
+sqlframe-3.4.1.dist-info/METADATA,sha256=AhBfRu7B6juTMVvsb9xiFk3pbpINYWEKKz8hASjyfgU,8639
+sqlframe-3.4.1.dist-info/WHEEL,sha256=G16H4A3IeoQmnOrYV4ueZGKSjhipXx8zc8nu9FGlvMA,92
+sqlframe-3.4.1.dist-info/top_level.txt,sha256=T0_RpoygaZSF6heeWwIDQgaP0varUdSK1pzjeJZRjM8,9
+sqlframe-3.4.1.dist-info/RECORD,,