def parse_identifier(item):
alias = item.get_alias()
sp_idx = item.token_next_by(t=Whitespace)[0] or len(item.tokens)
item_rev = Identifier(list(reversed(item.tokens[:sp_idx])))
name = item_rev._get_first_name(real_name=True)
alias = alias or name
dot_idx, _ = item_rev.token_next_by(m=(Punctuation, '.'))
if dot_idx is not None:
schema_name = item_rev._get_first_name(dot_idx, real_name=True)
dot_idx, _ = item_rev.token_next_by(m=(Punctuation, '.'),
idx=dot_idx)
if dot_idx is not None:
catalog_name = item_rev._get_first_name(dot_idx,
real_name=True)
else:
catalog_name = None
else:
schema_name = None
catalog_name = None
schema_quoted = schema_name and item.value[0] == '"'
if schema_name and not schema_quoted:
schema_name = schema_name.lower()
quote_count = item.value.count('"')
name_quoted = quote_count > 2 or (quote_count and not schema_quoted)
alias_quoted = alias and item.value[-1] == '"'
if alias_quoted or name_quoted and not alias and name.islower():
alias = '"' + (alias or name) + '"'
if name and not name_quoted and not name.islower():
if not alias:
alias = name
name = name.lower()
return catalog_name, schema_name, name, alias
def from_identifier(cls,
identifier: token_groups.Identifier,
position: int = 0) -> Column:
"""Create a column from an SQL identifier token.
Params:
-------
identifier: SQL token with column label.
Returns:
--------
A Column object based on the given identifier token.
"""
idx, identifier_name = identifier.token_next_by(
t=token_types.Name, i=token_groups.Function)
_, maybe_dot = identifier.token_next(idx, skip_ws=True, skip_cm=True)
if maybe_dot is None or not maybe_dot.match(token_types.Punctuation,
"."):
table_name = None
name = identifier_name.value
else:
table_name = identifier_name.value
idx, column_name_token = identifier.token_next_by(
t=token_types.Name, idx=idx)
# Fauna doesn't have an 'id' field, so we extract the ID value from the 'ref' included
# in query responses, but we still want to map the field name to aliases as with other
# fields for consistency when passing results to SQLAlchemy
name = "ref" if column_name_token.value == "id" else column_name_token.value
idx, as_keyword = identifier.token_next_by(m=(token_types.Keyword,
"AS"),
idx=idx)
if as_keyword is None:
alias = "id" if name == "ref" else name
else:
_, alias_identifier = identifier.token_next_by(
i=token_groups.Identifier, idx=idx)
alias = alias_identifier.value
function_name: typing.Optional[Function] = None
if re.match(COUNT_REGEX, name):
function_name = Function.COUNT
elif re.match(NOT_SUPPORTED_FUNCTION_REGEX, name):
raise exceptions.NotSupportedError(
"MIN, MAX, AVG, and SUM functions are not yet supported.")
column_params: ColumnParams = {
"table_name": table_name,
"name": name,
"alias": alias,
"function_name": function_name,
"position": position,
}
return Column(**column_params)
def get_identifier_parents(self):
if self.identifier is None:
return None, None
item_rev = Identifier(list(reversed(self.identifier.tokens)))
name = item_rev._get_first_name(real_name = True)
dot_idx, _ = item_rev.token_next_by(m=(Punctuation, '.'))
if dot_idx is not None:
schema_name = item_rev._get_first_name(dot_idx, real_name = True)
dot_idx, _ = item_rev.token_next_by(m=(Punctuation, '.'), idx = dot_idx)
if dot_idx is not None:
catalog_name = item_rev._get_first_name(dot_idx, real_name = True)
else:
catalog_name = None
else:
schema_name = None
catalog_name = None
return catalog_name, schema_name
def _extract_direction(
cls, identifier: token_groups.Identifier) -> OrderDirection:
_, direction_token = identifier.token_next_by(t=token_types.Keyword)
if direction_token is None:
PENULTIMATE_TOKEN = -2
# For some reason, when ordering by multiple columns with a direction keyword,
# sqlparse groups the final column with the direction in an Identifier token.
# There is an open issue (https://github.com/andialbrecht/sqlparse/issues/606),
# though without any response, so it seems to be a bug.
_, direction_identifier = identifier.token_next_by(
i=token_groups.Identifier, idx=PENULTIMATE_TOKEN)
if direction_identifier is not None:
_, direction_token = direction_identifier.token_next_by(
t=token_types.Keyword)
return (getattr(OrderDirection, direction_token.value)
if direction_token else None)
def from_identifier(cls, identifier: token_groups.Identifier) -> Table:
"""Extract table name from an SQL identifier.
Params:
-------
identifier: SQL token that contains the table's name.
Returns:
--------
A new Table object.
"""
idx, name = identifier.token_next_by(t=token_types.Name)
assert name is not None
idx, _ = identifier.token_next_by(m=(token_types.Keyword, "AS"),
idx=idx)
if idx is None:
return cls(name=name.value)
_, alias = identifier.token_next_by(i=token_groups.Identifier, idx=idx)
if alias is None:
return cls(name=name.value)
return cls(name=name.value, alias=alias.value)
def parse_identifier(item):
alias = item.get_alias()
sp_idx = item.token_next_by(t=Whitespace)[0] or len(item.tokens)
item_rev = Identifier(list(reversed(item.tokens[:sp_idx])))
name = item_rev._get_first_name(real_name=True)
alias = alias or name
dot_idx, _ = item_rev.token_next_by(m=(Punctuation, '.'))
if dot_idx is not None:
schema_name = item_rev._get_first_name(dot_idx, real_name=True)
dot_idx, _ = item_rev.token_next_by(m=(Punctuation, '.'),
idx=dot_idx)
if dot_idx is not None:
catalog_name = item_rev._get_first_name(dot_idx,
real_name=True)
else:
catalog_name = None
else:
schema_name = None
catalog_name = None
# TODO: this business below needs help
# for one we need to apply this logic to catalog_name
# then the logic around name_quoted = quote_count > 2 obviously
# doesn't work. Finally, quotechar needs to be customized
schema_quoted = schema_name and item.value[0] == '"'
if schema_name and not schema_quoted:
schema_name = schema_name.lower()
quote_count = item.value.count('"')
name_quoted = quote_count > 2 or (quote_count and not schema_quoted)
alias_quoted = alias and item.value[-1] == '"'
if alias_quoted or name_quoted and not alias and name.islower():
alias = '"' + (alias or name) + '"'
if name and not name_quoted and not name.islower():
if not alias:
alias = name
name = name.lower()
return catalog_name, schema_name, name, alias
