X-Git-Url: https://git.phdru.name/?p=sqlconvert.git;a=blobdiff_plain;f=sqlconvert%2Fprocess_mysql.py;h=1834855dc9b59aee5261cf065afcab1c58310fac;hp=3f2f6aa3c7dee8ad84f60c3a630050c8f3d413b6;hb=72abf4a136b1a2d164259a4ac300e6a5a4762432;hpb=ad12dfd9c03a4f6bbd03d22238b6399ab09962ed

diff --git a/sqlconvert/process_mysql.py b/sqlconvert/process_mysql.py
index 3f2f6aa..1834855 100644
--- a/sqlconvert/process_mysql.py
+++ b/sqlconvert/process_mysql.py
@@ -1,6 +1,8 @@
 
-from sqlparse.sql import Comment
+from sqlparse.sql import Comment, Function, Identifier, Parenthesis, \
+    Statement, Token
 from sqlparse import tokens as T
+from .process_tokens import escape_strings, is_comment_or_space
 
 
 def _is_directive_token(token):
@@ -26,8 +28,8 @@ def is_directive_statement(statement):
     return True
 
 
-def remove_directives(statement):
-    """Remove /*! directives */ from the first-level"""
+def remove_directive_tokens(statement):
+    """Remove /\*! directives \*/ from the first-level"""
     new_tokens = []
     for token in statement.tokens:
         if _is_directive_token(token):
@@ -49,6 +51,100 @@ def requote_names(token_list):
                 token.normalized = token.value = '"%s"' % value
 
 
-def process_statement(statement):
-    remove_directives(statement)
+def unescape_strings(token_list):
+    """Unescape strings"""
+    for token in token_list.flatten():
+        if token.ttype is T.String.Single:
+            value = token.value
+            for orig, repl in (
+                ('\\"', '"'),
+                ("\\'", "'"),
+                ("''", "'"),
+                ('\\b', '\b'),
+                ('\\n', '\n'),
+                ('\\r', '\r'),
+                ('\\t', '\t'),
+                ('\\\032', '\032'),
+                ('\\\\', '\\'),
+            ):
+                value = value.replace(orig, repl)
+            token.normalized = token.value = value
+
+
+def is_insert(statement):
+    for token in statement.tokens:
+        if is_comment_or_space(token):
+            continue
+        return (token.ttype is T.DML) and (token.normalized == 'INSERT')
+
+
+def split_ext_insert(statement):
+    """Split extended INSERT into multiple standard INSERTs"""
+    insert_tokens = []
+    values_tokens = []
+    end_tokens = []
+    expected = 'INSERT'
+    for token in statement.tokens:
+        if is_comment_or_space(token):
+            if expected == 'END':
+                end_tokens.append(token)
+            else:
+                insert_tokens.append(token)
+            continue
+        elif expected == 'INSERT':
+            if (token.ttype is T.DML) and (token.normalized == 'INSERT'):
+                insert_tokens.append(token)
+                expected = 'INTO'
+                continue
+        elif expected == 'INTO':
+            if (token.ttype is T.Keyword) and (token.normalized == 'INTO'):
+                insert_tokens.append(token)
+                expected = 'TABLE_NAME'
+                continue
+        elif expected == 'TABLE_NAME':
+            if isinstance(token, (Function, Identifier)):
+                insert_tokens.append(token)
+                expected = 'VALUES'
+                continue
+        elif expected == 'VALUES':
+            if (token.ttype is T.Keyword) and (token.normalized == 'VALUES'):
+                insert_tokens.append(token)
+                expected = 'VALUES_OR_SEMICOLON'
+                continue
+        elif expected == 'VALUES_OR_SEMICOLON':
+            if isinstance(token, Parenthesis):
+                values_tokens.append(token)
+                continue
+            elif token.ttype is T.Punctuation:
+                if token.value == ',':
+                    continue
+                elif token.value == ';':
+                    end_tokens.append(token)
+                    expected = 'END'
+                    continue
+        raise ValueError(
+            'SQL syntax error: expected "%s", got %s "%s"' % (
+                expected, token.ttype, token.normalized))
+    new_line = Token(T.Newline, '\n')
+    new_lines = [new_line]  # Insert newlines between split statements
+    for i, values in enumerate(values_tokens):
+        if i == len(values_tokens) - 1:  # Last but one statement
+            # Insert newlines only between split statements but not after
+            new_lines = []
+        # The statement sets `parent` attribute of the every token to self
+        # but we don't care.
+        statement = Statement(insert_tokens + [values] +
+                              end_tokens + new_lines)
+        yield statement
+
+
+def process_statement(statement, quoting_style='sqlite'):
     requote_names(statement)
+    unescape_strings(statement)
+    remove_directive_tokens(statement)
+    escape_strings(statement, quoting_style)
+    if is_insert(statement):
+        for statement in split_ext_insert(statement):
+            yield statement
+    else:
+        yield statement