Commit 600f7f22 by Sartika Aritonang

Upload New File

parent 4ee3f39e
# -*- coding: utf-8 -*-
#
# Copyright (C) 2009-2018 the sqlparse authors and contributors
# <see AUTHORS file>
#
# This module is part of python-sqlparse and is released under
# the BSD License: https://opensource.org/licenses/BSD-3-Clause
from sqlparse import sql, tokens as T
from sqlparse.utils import split_unquoted_newlines
class StripCommentsFilter(object):
@staticmethod
def _process(tlist):
def get_next_comment():
# TODO(andi) Comment types should be unified, see related issue38
return tlist.token_next_by(i=sql.Comment, t=T.Comment)
tidx, token = get_next_comment()
while token:
pidx, prev_ = tlist.token_prev(tidx, skip_ws=False)
nidx, next_ = tlist.token_next(tidx, skip_ws=False)
# Replace by whitespace if prev and next exist and if they're not
# whitespaces. This doesn't apply if prev or next is a parenthesis.
if (prev_ is None or next_ is None
or prev_.is_whitespace or prev_.match(T.Punctuation, '(')
or next_.is_whitespace or next_.match(T.Punctuation, ')')):
# Insert a whitespace to ensure the following SQL produces
# a valid SQL (see #425). For example:
#
# Before: select a--comment\nfrom foo
# After: select a from foo
if prev_ is not None and next_ is None:
tlist.tokens.insert(tidx, sql.Token(T.Whitespace, ' '))
tlist.tokens.remove(token)
else:
tlist.tokens[tidx] = sql.Token(T.Whitespace, ' ')
tidx, token = get_next_comment()
def process(self, stmt):
[self.process(sgroup) for sgroup in stmt.get_sublists()]
StripCommentsFilter._process(stmt)
return stmt
class StripWhitespaceFilter(object):
def _stripws(self, tlist):
func_name = '_stripws_{cls}'.format(cls=type(tlist).__name__)
func = getattr(self, func_name.lower(), self._stripws_default)
func(tlist)
@staticmethod
def _stripws_default(tlist):
last_was_ws = False
is_first_char = True
for token in tlist.tokens:
if token.is_whitespace:
token.value = '' if last_was_ws or is_first_char else ' '
last_was_ws = token.is_whitespace
is_first_char = False
def _stripws_identifierlist(self, tlist):
# Removes newlines before commas, see issue140
last_nl = None
for token in list(tlist.tokens):
if last_nl and token.ttype is T.Punctuation and token.value == ',':
tlist.tokens.remove(last_nl)
last_nl = token if token.is_whitespace else None
# next_ = tlist.token_next(token, skip_ws=False)
# if (next_ and not next_.is_whitespace and
# token.ttype is T.Punctuation and token.value == ','):
# tlist.insert_after(token, sql.Token(T.Whitespace, ' '))
return self._stripws_default(tlist)
def _stripws_parenthesis(self, tlist):
while tlist.tokens[1].is_whitespace:
tlist.tokens.pop(1)
while tlist.tokens[-2].is_whitespace:
tlist.tokens.pop(-2)
self._stripws_default(tlist)
def process(self, stmt, depth=0):
[self.process(sgroup, depth + 1) for sgroup in stmt.get_sublists()]
self._stripws(stmt)
if depth == 0 and stmt.tokens and stmt.tokens[-1].is_whitespace:
stmt.tokens.pop(-1)
return stmt
class SpacesAroundOperatorsFilter(object):
@staticmethod
def _process(tlist):
ttypes = (T.Operator, T.Comparison)
tidx, token = tlist.token_next_by(t=ttypes)
while token:
nidx, next_ = tlist.token_next(tidx, skip_ws=False)
if next_ and next_.ttype != T.Whitespace:
tlist.insert_after(tidx, sql.Token(T.Whitespace, ' '))
pidx, prev_ = tlist.token_prev(tidx, skip_ws=False)
if prev_ and prev_.ttype != T.Whitespace:
tlist.insert_before(tidx, sql.Token(T.Whitespace, ' '))
tidx += 1 # has to shift since token inserted before it
# assert tlist.token_index(token) == tidx
tidx, token = tlist.token_next_by(t=ttypes, idx=tidx)
def process(self, stmt):
[self.process(sgroup) for sgroup in stmt.get_sublists()]
SpacesAroundOperatorsFilter._process(stmt)
return stmt
# ---------------------------
# postprocess
class SerializerUnicode(object):
@staticmethod
def process(stmt):
lines = split_unquoted_newlines(stmt)
return '\n'.join(line.rstrip() for line in lines)
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment