jupyter-duckdb 1.2.0.5__tar.gz → 1.2.7__tar.gz

{jupyter_duckdb-1.2.0.5 → jupyter_duckdb-1.2.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: jupyter-duckdb
-Version: 1.2.0.5
+Version: 1.2.7
 Summary: a basic wrapper kernel for DuckDB
 Home-page: https://github.com/erictroebs/jupyter-duckdb
 Author: Eric Tröbs

{jupyter_duckdb-1.2.0.5 → jupyter_duckdb-1.2.7}/src/duckdb_kernel/db/implementation/duckdb/Connection.py

@@ -4,6 +4,7 @@ import duckdb
 
 from ... import DatabaseError, Column, Constraint, ForeignKey, Table
 from ...Connection import Connection as Base
+from ...error import EmptyResultError
 
 
 class Connection(Base):
@@ -36,7 +37,10 @@ class Connection(Base):
                     raise e
 
             # get rows
-            rows = cursor.fetchall()
+            try:
+                rows = cursor.fetchall()
+            except duckdb.InvalidInputException as e:
+                raise EmptyResultError(str(e))
 
             # get columns
             if cursor.description is None:

{jupyter_duckdb-1.2.0.5 → jupyter_duckdb-1.2.7}/src/duckdb_kernel/db/implementation/sqlite/Connection.py

@@ -4,6 +4,7 @@ from typing import Dict, List, Tuple, Any
 
 from ... import DatabaseError, Column, Constraint, ForeignKey, Table
 from ...Connection import Connection as Base
+from ...error import EmptyResultError
 
 
 class Connection(Base):
@@ -41,7 +42,7 @@ class Connection(Base):
 
             # get columns
             if cursor.description is None:
-                columns = []
+                raise EmptyResultError()
             else:
                 columns = [e[0] for e in cursor.description]
 

{jupyter_duckdb-1.2.0.5 → jupyter_duckdb-1.2.7}/src/duckdb_kernel/kernel.py

@@ -14,6 +14,7 @@ from .db.error import *
 from .magics import *
 from .parser import RAParser, DCParser, ParserError
 from .util.ResultSetComparator import ResultSetComparator
+from .util.SQL import SQL_KEYWORDS
 from .util.TestError import TestError
 from .util.formatting import row_count, rows_table, wrap_image
 from .visualization import *
@@ -46,11 +47,12 @@ class DuckDBKernel(Kernel):
             MagicCommand('query_max_rows').arg('count').on(self._query_max_rows_magic),
             MagicCommand('schema').flag('td').opt('only').on(self._schema_magic),
             MagicCommand('store').arg('file').flag('noheader').result(True).on(self._store_magic),
-            MagicCommand('ra').disable('dc', 'auto_parser').flag('analyze').code(True).on(self._ra_magic),
+            MagicCommand('sql').disable('ra', 'dc', 'auto_parser'),
+            MagicCommand('ra').disable('sql', 'dc', 'auto_parser').flag('analyze').code(True).on(self._ra_magic),
             MagicCommand('all_ra').arg('value', '1').on(self._all_ra_magic),
-            MagicCommand('dc').disable('ra', 'auto_parser').code(True).on(self._dc_magic),
+            MagicCommand('dc').disable('sql', 'ra', 'auto_parser').code(True).on(self._dc_magic),
             MagicCommand('all_dc').arg('value', '1').on(self._all_dc_magic),
-            MagicCommand('auto_parser').disable('ra', 'dc').code(True).on(self._auto_parser_magic),
+            MagicCommand('auto_parser').disable('sql', 'ra', 'dc').code(True).on(self._auto_parser_magic),
             MagicCommand('guess_parser').arg('value', '1').on(self._guess_parser_magic),
         )
 
@@ -301,6 +303,11 @@ class DuckDBKernel(Kernel):
                 self.print(f'loaded tests from {with_tests}\n')
 
     def _test_magic(self, silent: bool, result_columns: List[str], result: List[List], name: str):
+        # If the query was empty, result_columns and result may be None.
+        if result_columns is None or result is None:
+            self.print_data(wrap_image(False, 'Statement did not return data.'))
+            return
+
         # Testing makes no sense if there is no output.
         if silent:
             return
@@ -550,14 +557,22 @@ class DuckDBKernel(Kernel):
             self._magics['auto_parser'].default(False)
 
     def _auto_parser_magic(self, silent: bool, code: str):
+        # do not handle statements starting with SQL keywords
+        first_word = code.strip().split(maxsplit=1)
+        if len(first_word) > 0:
+            if first_word[0].upper() in SQL_KEYWORDS:
+                return
+
+        # try to parse DC
         try:
-            return self._ra_magic(silent, code, analyze=False)
+            return self._dc_magic(silent, code)
         except ParserError as e:
             if e.depth > 0:
                 raise e
 
+        # try to parse RA
         try:
-            return self._dc_magic(silent, code)
+            return self._ra_magic(silent, code, analyze=False)
         except ParserError as e:
             if e.depth > 0:
                 raise e

{jupyter_duckdb-1.2.0.5 → jupyter_duckdb-1.2.7}/src/duckdb_kernel/parser/RAParser.py

@@ -19,10 +19,14 @@ class RAParser:
             tokens = tuple(Tokenizer.tokenize(tokens[0]))
 
         # binary operators
-        for operator in RA_BINARY_OPERATORS:
+        for operator_symbols in RA_BINARY_SYMBOLS:
             # iterate tokens and match symbol
             for i in range(1, len(tokens) + 1):
-                if tokens[-i].lower() in operator.symbols():
+                lower_token = tokens[-i].lower()
+
+                if lower_token in operator_symbols:
+                    operator = operator_symbols[lower_token]
+
                     # raise error if left or right operand missing
                     if i == 1:
                         raise RAParserError(f'right operand missing after {tokens[-i]}', depth)

{jupyter_duckdb-1.2.0.5 → jupyter_duckdb-1.2.7}/src/duckdb_kernel/parser/elements/__init__.py

@@ -9,7 +9,7 @@ from .RAOperator import RAOperator
 from .RABinaryOperator import RABinaryOperator
 from .RAUnaryOperator import RAUnaryOperator
 from .RAOperand import RAOperand
-from .binary import RA_BINARY_OPERATORS
+from .binary import RA_BINARY_OPERATORS, RA_BINARY_SYMBOLS
 from .unary import RA_UNARY_OPERATORS
 
 from .DCOperand import DCOperand

jupyter_duckdb-1.2.7/src/duckdb_kernel/parser/elements/binary/FullOuterJoin.py

@@ -0,0 +1,30 @@
+from typing import Tuple, Dict
+
+from duckdb_kernel.db import Table
+from ..RABinaryOperator import RABinaryOperator
+from ...util.RenamableColumnList import RenamableColumnList
+
+
+class FullOuterJoin(RABinaryOperator):
+    @staticmethod
+    def symbols() -> Tuple[str, ...]:
+        return chr(10199), 'fjoin', 'ojoin'
+
+    def to_sql(self, tables: Dict[str, Table]) -> Tuple[str, RenamableColumnList]:
+        # execute subqueries
+        lq, lcols = self.left.to_sql(tables)
+        rq, rcols = self.right.to_sql(tables)
+
+        # find matching columns
+        join_cols, all_cols = lcols.intersect(rcols)
+
+        replacements = {c1: c2 for c1, c2 in join_cols}
+        select_cols = [
+            f'COALESCE({c.current_name}, {replacements.get(c).current_name})' if c in replacements else c.current_name
+            for c in all_cols]
+        select_clause = ', '.join(select_cols)
+
+        on_clause = ' AND '.join(f'{l.current_name} = {r.current_name}' for l, r in join_cols)
+
+        # create sql
+        return f'SELECT {select_clause} FROM ({lq}) {self._name()} FULL OUTER JOIN ({rq}) {self._name()} ON {on_clause}', all_cols

jupyter_duckdb-1.2.7/src/duckdb_kernel/parser/elements/binary/LeftOuterJoin.py

@@ -0,0 +1,24 @@
+from typing import Tuple, Dict
+
+from duckdb_kernel.db import Table
+from ..RABinaryOperator import RABinaryOperator
+from ...util.RenamableColumnList import RenamableColumnList
+
+
+class LeftOuterJoin(RABinaryOperator):
+    @staticmethod
+    def symbols() -> Tuple[str, ...]:
+        return chr(10197), 'ljoin'
+
+    def to_sql(self, tables: Dict[str, Table]) -> Tuple[str, RenamableColumnList]:
+        # execute subqueries
+        lq, lcols = self.left.to_sql(tables)
+        rq, rcols = self.right.to_sql(tables)
+
+        # find matching columns
+        join_cols, all_cols = lcols.intersect(rcols)
+
+        on_clause = ' AND '.join(f'{l.current_name} = {r.current_name}' for l, r in join_cols)
+
+        # create sql
+        return f'SELECT {all_cols.list} FROM ({lq}) {self._name()} LEFT OUTER JOIN ({rq}) {self._name()} ON {on_clause}', all_cols

jupyter_duckdb-1.2.7/src/duckdb_kernel/parser/elements/binary/RightOuterJoin.py

@@ -0,0 +1,24 @@
+from typing import Tuple, Dict
+
+from duckdb_kernel.db import Table
+from ..RABinaryOperator import RABinaryOperator
+from ...util.RenamableColumnList import RenamableColumnList
+
+
+class RightOuterJoin(RABinaryOperator):
+    @staticmethod
+    def symbols() -> Tuple[str, ...]:
+        return chr(10198), 'rjoin'
+
+    def to_sql(self, tables: Dict[str, Table]) -> Tuple[str, RenamableColumnList]:
+        # execute subqueries
+        lq, lcols = self.left.to_sql(tables)
+        rq, rcols = self.right.to_sql(tables)
+
+        # find matching columns
+        join_cols, all_cols = lcols.intersect(rcols, prefer_right=True)
+
+        on_clause = ' AND '.join(f'{l.current_name} = {r.current_name}' for l, r in join_cols)
+
+        # create sql
+        return f'SELECT {all_cols.list} FROM ({lq}) {self._name()} RIGHT OUTER JOIN ({rq}) {self._name()} ON {on_clause}', all_cols

{jupyter_duckdb-1.2.0.5 → jupyter_duckdb-1.2.7}/src/duckdb_kernel/parser/elements/binary/__init__.py

@@ -2,6 +2,9 @@ from .Cross import Cross
 from .Difference import Difference
 from .Intersection import Intersection
 from .Join import Join
+from .LeftOuterJoin import LeftOuterJoin
+from .RightOuterJoin import RightOuterJoin
+from .FullOuterJoin import FullOuterJoin
 from .Union import Union
 
 from .Add import Add
@@ -30,12 +33,22 @@ LOGIC_BINARY_OPERATORS = sorted([
 ], key=lambda x: x.order, reverse=True)
 
 RA_BINARY_OPERATORS = [
-    Difference,
-    Union,
-    Intersection,
-    Join,
-    Cross,
-    Division
+    [Difference],
+    [Union],
+    [Intersection],
+    [Join],
+    [LeftOuterJoin, RightOuterJoin, FullOuterJoin],
+    [Cross],
+    [Division]
+]
+
+RA_BINARY_SYMBOLS = [
+    {
+        symbol: operator
+        for operator in level
+        for symbol in operator.symbols()
+    }
+    for level in RA_BINARY_OPERATORS
 ]
 
 DC_SET = ConditionalSet

{jupyter_duckdb-1.2.0.5 → jupyter_duckdb-1.2.7}/src/duckdb_kernel/parser/util/RenamableColumnList.py

@@ -73,18 +73,26 @@ class RenamableColumnList(list[RenamableColumn]):
 
         return RenamableColumnList(cols.values())
 
-    def intersect(self, other: 'RenamableColumnList') \
+    def intersect(self, other: 'RenamableColumnList', prefer_right: bool = False) \
             -> Tuple[List[Tuple[RenamableColumn, RenamableColumn]], 'RenamableColumnList']:
         self_cols: Dict[str, RenamableColumn] = {col.name: col for col in self}
         other_cols: Dict[str, RenamableColumn] = {col.name: col for col in other}
 
+        replacements: Dict[RenamableColumn, RenamableColumn] = {}
         intersection: List[Tuple[RenamableColumn, RenamableColumn]] = []
+
         for name in tuple(self_cols.keys()):
             if name in other_cols:
+                if prefer_right:
+                    replacements[self_cols[name]] = other_cols[name]
+
                 intersection.append((self_cols[name], other_cols[name]))
                 del other_cols[name]
 
         # if len(intersection) == 0:
         #     raise AssertionError('no common attributes found for join')
 
-        return intersection, RenamableColumnList(self_cols.values(), other_cols.values())
+        return intersection, RenamableColumnList(
+            (replacements.get(x, x) for x in self_cols.values()),
+            other_cols.values()
+        )

jupyter_duckdb-1.2.7/src/duckdb_kernel/util/SQL.py

@@ -0,0 +1,6 @@
+# see https://www.postgresql.org/docs/current/sql-commands.html
+SQL_KEYWORDS = ('ABORT', 'ALTER', 'ANALYZE', 'BEGIN', 'CALL', 'CHECKPOINT', 'CLOSE', 'CLUSTER', 'COMMENT', 'COMMIT',
+                'COPY', 'CREATE', 'DEALLOCATE', 'DECLARE', 'DELETE', 'DISCARD', 'DO', 'DROP', 'END', 'EXECUTE',
+                'EXPLAIN', 'FETCH', 'GRANT', 'IMPORT', 'INSERT', 'LISTEN', 'LOAD', 'LOCK', 'MERGE', 'MOVE', 'NOTIFY',
+                'PREPARE', 'REASSIGN', 'REFRESH', 'REINDEX', 'RELEASE', 'RESET', 'REVOKE', 'ROLLBACK', 'SAVEPOINT',
+                'SECURITY', 'SELECT', 'SET', 'SHOW', 'START', 'TRUNCATE', 'UNLISTEN', 'UPDATE', 'VACUUM', 'VALUES')

{jupyter_duckdb-1.2.0.5 → jupyter_duckdb-1.2.7}/src/jupyter_duckdb.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: jupyter-duckdb
-Version: 1.2.0.5
+Version: 1.2.7
 Summary: a basic wrapper kernel for DuckDB
 Home-page: https://github.com/erictroebs/jupyter-duckdb
 Author: Eric Tröbs

{jupyter_duckdb-1.2.0.5 → jupyter_duckdb-1.2.7}/src/jupyter_duckdb.egg-info/SOURCES.txt

@@ -50,15 +50,18 @@ src/duckdb_kernel/parser/elements/binary/Difference.py
 src/duckdb_kernel/parser/elements/binary/Divide.py
 src/duckdb_kernel/parser/elements/binary/Division.py
 src/duckdb_kernel/parser/elements/binary/Equal.py
+src/duckdb_kernel/parser/elements/binary/FullOuterJoin.py
 src/duckdb_kernel/parser/elements/binary/GreaterThan.py
 src/duckdb_kernel/parser/elements/binary/GreaterThanEqual.py
 src/duckdb_kernel/parser/elements/binary/Intersection.py
 src/duckdb_kernel/parser/elements/binary/Join.py
+src/duckdb_kernel/parser/elements/binary/LeftOuterJoin.py
 src/duckdb_kernel/parser/elements/binary/LessThan.py
 src/duckdb_kernel/parser/elements/binary/LessThanEqual.py
 src/duckdb_kernel/parser/elements/binary/Minus.py
 src/duckdb_kernel/parser/elements/binary/Multiply.py
 src/duckdb_kernel/parser/elements/binary/Or.py
+src/duckdb_kernel/parser/elements/binary/RightOuterJoin.py
 src/duckdb_kernel/parser/elements/binary/Unequal.py
 src/duckdb_kernel/parser/elements/binary/Union.py
 src/duckdb_kernel/parser/elements/binary/__init__.py
@@ -74,6 +77,7 @@ src/duckdb_kernel/parser/util/RenamableColumn.py
 src/duckdb_kernel/parser/util/RenamableColumnList.py
 src/duckdb_kernel/parser/util/__init__.py
 src/duckdb_kernel/util/ResultSetComparator.py
+src/duckdb_kernel/util/SQL.py
 src/duckdb_kernel/util/TestError.py
 src/duckdb_kernel/util/__init__.py
 src/duckdb_kernel/util/formatting.py
@@ -88,4 +92,5 @@ src/jupyter_duckdb.egg-info/requires.txt
 src/jupyter_duckdb.egg-info/top_level.txt
 test/test_dc.py
 test/test_ra.py
-test/test_result_comparison.py
+test/test_result_comparison.py
+test/test_sql.py