Shortcuts on this page
r m x toggle line displays
j k next/prev highlighted chunk
0 (zero) top of page
1 (one) first highlighted chunk
1#
2# Copyright (C) 2009-2020 the sqlparse authors and contributors
3# <see AUTHORS file>
4#
5# This module is part of python-sqlparse and is released under
6# the BSD License: https://opensource.org/licenses/BSD-3-Clause
8import re
10from sqlparse import sql, tokens as T
11from sqlparse.utils import split_unquoted_newlines
14class StripCommentsFilter:
16 @staticmethod
17 def _process(tlist):
18 def get_next_comment(idx=-1):
19 # TODO(andi) Comment types should be unified, see related issue38
20 return tlist.token_next_by(i=sql.Comment, t=T.Comment, idx=idx)
22 def _get_insert_token(token):
23 """Returns either a whitespace or the line breaks from token."""
24 # See issue484 why line breaks should be preserved.
25 # Note: The actual value for a line break is replaced by \n
26 # in SerializerUnicode which will be executed in the
27 # postprocessing state.
28 m = re.search(r'([\r\n]+) *$', token.value)
29 if m is not None:
30 return sql.Token(T.Whitespace.Newline, m.groups()[0])
31 else:
32 return sql.Token(T.Whitespace, ' ')
34 sql_hints = (T.Comment.Multiline.Hint, T.Comment.Single.Hint)
35 tidx, token = get_next_comment()
36 while token:
37 # skipping token remove if token is a SQL-Hint. issue262
38 is_sql_hint = False
39 if token.ttype in sql_hints:
40 is_sql_hint = True
41 elif isinstance(token, sql.Comment):
42 comment_tokens = token.tokens
43 if len(comment_tokens) > 0:
44 if comment_tokens[0].ttype in sql_hints:
45 is_sql_hint = True
47 if is_sql_hint:
48 # using current index as start index to search next token for
49 # preventing infinite loop in cases when token type is a
50 # "SQL-Hint" and has to be skipped
51 tidx, token = get_next_comment(idx=tidx)
52 continue
54 pidx, prev_ = tlist.token_prev(tidx, skip_ws=False)
55 nidx, next_ = tlist.token_next(tidx, skip_ws=False)
56 # Replace by whitespace if prev and next exist and if they're not
57 # whitespaces. This doesn't apply if prev or next is a parenthesis.
58 if (
59 prev_ is None or next_ is None
60 or prev_.is_whitespace or prev_.match(T.Punctuation, '(')
61 or next_.is_whitespace or next_.match(T.Punctuation, ')')
62 ):
63 # Insert a whitespace to ensure the following SQL produces
64 # a valid SQL (see #425).
65 if prev_ is not None and not prev_.match(T.Punctuation, '('):
66 tlist.tokens.insert(tidx, _get_insert_token(token))
67 tlist.tokens.remove(token)
68 else:
69 tlist.tokens[tidx] = _get_insert_token(token)
71 # using current index as start index to search next token for
72 # preventing infinite loop in cases when token type is a
73 # "SQL-Hint" and has to be skipped
74 tidx, token = get_next_comment(idx=tidx)
76 def process(self, stmt):
77 [self.process(sgroup) for sgroup in stmt.get_sublists()]
78 StripCommentsFilter._process(stmt)
79 return stmt
82class StripWhitespaceFilter:
83 def _stripws(self, tlist):
84 func_name = f'_stripws_{type(tlist).__name__}'
85 func = getattr(self, func_name.lower(), self._stripws_default)
86 func(tlist)
88 @staticmethod
89 def _stripws_default(tlist):
90 last_was_ws = False
91 is_first_char = True
92 for token in tlist.tokens:
93 if token.is_whitespace:
94 token.value = '' if last_was_ws or is_first_char else ' '
95 last_was_ws = token.is_whitespace
96 is_first_char = False
98 def _stripws_identifierlist(self, tlist):
99 # Removes newlines before commas, see issue140
100 last_nl = None
101 for token in list(tlist.tokens):
102 if last_nl and token.ttype is T.Punctuation and token.value == ',':
103 tlist.tokens.remove(last_nl)
104 last_nl = token if token.is_whitespace else None
106 # next_ = tlist.token_next(token, skip_ws=False)
107 # if (next_ and not next_.is_whitespace and
108 # token.ttype is T.Punctuation and token.value == ','):
109 # tlist.insert_after(token, sql.Token(T.Whitespace, ' '))
110 return self._stripws_default(tlist)
112 def _stripws_parenthesis(self, tlist):
113 while tlist.tokens[1].is_whitespace:
114 tlist.tokens.pop(1)
115 while tlist.tokens[-2].is_whitespace:
116 tlist.tokens.pop(-2)
117 if tlist.tokens[-2].is_group:
118 # save to remove the last whitespace
119 while tlist.tokens[-2].tokens[-1].is_whitespace:
120 tlist.tokens[-2].tokens.pop(-1)
121 self._stripws_default(tlist)
123 def process(self, stmt, depth=0):
124 [self.process(sgroup, depth + 1) for sgroup in stmt.get_sublists()]
125 self._stripws(stmt)
126 if depth == 0 and stmt.tokens and stmt.tokens[-1].is_whitespace:
127 stmt.tokens.pop(-1)
128 return stmt
131class SpacesAroundOperatorsFilter:
132 @staticmethod
133 def _process(tlist):
135 ttypes = (T.Operator, T.Comparison)
136 tidx, token = tlist.token_next_by(t=ttypes)
137 while token:
138 nidx, next_ = tlist.token_next(tidx, skip_ws=False)
139 if next_ and next_.ttype != T.Whitespace:
140 tlist.insert_after(tidx, sql.Token(T.Whitespace, ' '))
142 pidx, prev_ = tlist.token_prev(tidx, skip_ws=False)
143 if prev_ and prev_.ttype != T.Whitespace:
144 tlist.insert_before(tidx, sql.Token(T.Whitespace, ' '))
145 tidx += 1 # has to shift since token inserted before it
147 # assert tlist.token_index(token) == tidx
148 tidx, token = tlist.token_next_by(t=ttypes, idx=tidx)
150 def process(self, stmt):
151 [self.process(sgroup) for sgroup in stmt.get_sublists()]
152 SpacesAroundOperatorsFilter._process(stmt)
153 return stmt
156class StripTrailingSemicolonFilter:
158 def process(self, stmt):
159 while stmt.tokens and (stmt.tokens[-1].is_whitespace
160 or stmt.tokens[-1].value == ';'):
161 stmt.tokens.pop()
162 return stmt
165# ---------------------------
166# postprocess
168class SerializerUnicode:
169 @staticmethod
170 def process(stmt):
171 lines = split_unquoted_newlines(stmt)
172 return '\n'.join(line.rstrip() for line in lines)