Remove double quotes from filenames

[mihuman]

This pull request addresses an issue with table names that include double quotes.

Previously, the split_sql_file function in pg_split_schema_dump.py would generate filenames that included double quotes if the table name in the SQL dump file was double-quoted (using a reserved keyword as table name, for example 'user'). This could cause an error when trying to write the file, as double quotes are not allowed in filenames on Windows. Example:

OSError: [Errno 22] Invalid argument: '.\\split-dump\\my_schema.TABLE_"user".ACL'

The change in this pull request modifies the split_sql_file function to remove double quotes from table names when generating filenames, and adds a test case.

[akaihola]

Thanks @mihuman, this looks great!

Could you just fix the code formatting to conform to Black? I'm running Darker to enforce that on modified lines of each PR, and this is what we now get from the build run:

--- pgtricks/pg_split_schema_dump.py	2024-04-13 07:59:41.699312 +0000
+++ pgtricks/pg_split_schema_dump.py	2024-04-13 07:59:55.328318 +0000
@@ -62,11 +62,11 @@
                 "{chunk}\n"
                 "{divider}".format(sqlpath=sqlpath, chunk=part, divider=77 * '=')
             )
             print(part)
             continue
-        name = match.group(1).replace(" ", "_").replace('"', '')
+        name = match.group(1).replace(" ", "_").replace('"', "")
         type_ = match.group(2).replace(" ", "_")
         schema = match.group(3)
         if schema == "-":
             schema = "no_schema"
         if name == "id":
--- pgtricks/tests/test_pg_split_schema_dump.py	2024-04-13 07:59:41.699312 +0000
+++ pgtricks/tests/test_pg_split_schema_dump.py	2024-04-13 07:59:55.389746 +0000
@@ -150,38 +150,39 @@
 
         (information for the unidentified content goes here)
         ============================================================================="""
     )
 
+
 def test_split_sql_file_with_quotes_in_name(tmpdir):
-    target_directory = tmpdir / 'target'
-    sqlpath = tmpdir / 'test.sql'
+    target_directory = tmpdir / "target"
+    sqlpath = tmpdir / "test.sql"
     sqlpath.write(
         dedent(
-            '''
+            """
 
             --
             -- Name: "user"; Type: TABLE; Schema: public; Owner:
             --
 
             (information for user table goes here)
-            '''
+            """
         )
     )
 
     split_sql_file(str(sqlpath), str(target_directory))
 
     assert {path.basename for path in target_directory.listdir()} == {
-        'public.user.TABLE'
+        "public.user.TABLE"
     }
-    assert (target_directory / 'public.user.TABLE').readlines(cr=False) == [
-        'SET search_path = public;',
-        '',
-        '',
-        '',
-        '--',
+    assert (target_directory / "public.user.TABLE").readlines(cr=False) == [
+        "SET search_path = public;",
+        "",
+        "",
+        "",
+        "--",
         '-- Name: "user"; Type: TABLE; Schema: public; Owner:',
-        '--',
-        '',
-        '(information for user table goes here)',
-        '',
+        "--",
+        "",
+        "(information for user table goes here)",
+        "",
     ]

pgtricks/tests/test_pg_split_schema_dump.py:155:1: E302 expected 2 blank lines, found 1 [flake8]

[mihuman]

Hi @akaihola , sorry about that. The formatting should be fixed now.