/usr/lib/python3/dist-packages/sqlparse/engine/statement_splitter.py is in python3-sqlparse 0.2.4-0.1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 | # -*- coding: utf-8 -*-
#
# Copyright (C) 2016 Andi Albrecht, albrecht.andi@gmail.com
#
# 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
class StatementSplitter(object):
"""Filter that split stream at individual statements"""
def __init__(self):
self._reset()
def _reset(self):
"""Set the filter attributes to its default values"""
self._in_declare = False
self._is_create = False
self._begin_depth = 0
self.consume_ws = False
self.tokens = []
self.level = 0
def _change_splitlevel(self, ttype, value):
"""Get the new split level (increase, decrease or remain equal)"""
# ANSI
# if normal token return
# wouldn't parenthesis increase/decrease a level?
# no, inside a paranthesis can't start new statement
if ttype not in T.Keyword:
return 0
# Everything after here is ttype = T.Keyword
# Also to note, once entered an If statement you are done and basically
# returning
unified = value.upper()
# three keywords begin with CREATE, but only one of them is DDL
# DDL Create though can contain more words such as "or replace"
if ttype is T.Keyword.DDL and unified.startswith('CREATE'):
self._is_create = True
return 0
# can have nested declare inside of being...
if unified == 'DECLARE' and self._is_create and self._begin_depth == 0:
self._in_declare = True
return 1
if unified == 'BEGIN':
self._begin_depth += 1
if self._is_create:
# FIXME(andi): This makes no sense.
return 1
return 0
# Should this respect a preceeding BEGIN?
# In CASE ... WHEN ... END this results in a split level -1.
# Would having multiple CASE WHEN END and a Assigment Operator
# cause the statement to cut off prematurely?
if unified == 'END':
self._begin_depth = max(0, self._begin_depth - 1)
return -1
if (unified in ('IF', 'FOR', 'WHILE') and
self._is_create and self._begin_depth > 0):
return 1
if unified in ('END IF', 'END FOR', 'END WHILE'):
return -1
# Default
return 0
def process(self, stream):
"""Process the stream"""
EOS_TTYPE = T.Whitespace, T.Comment.Single
# Run over all stream tokens
for ttype, value in stream:
# Yield token if we finished a statement and there's no whitespaces
# It will count newline token as a non whitespace. In this context
# whitespace ignores newlines.
# why don't multi line comments also count?
if self.consume_ws and ttype not in EOS_TTYPE:
yield sql.Statement(self.tokens)
# Reset filter and prepare to process next statement
self._reset()
# Change current split level (increase, decrease or remain equal)
self.level += self._change_splitlevel(ttype, value)
# Append the token to the current statement
self.tokens.append(sql.Token(ttype, value))
# Check if we get the end of a statement
if self.level <= 0 and ttype is T.Punctuation and value == ';':
self.consume_ws = True
# Yield pending statement (if any)
if self.tokens:
yield sql.Statement(self.tokens)
|