def _enum_rows(self,
fh: IO[str],
_slice: slice = None) -> Iterator[Tuple[int, Row]]:
"""Enumerate on-disk and in-memory rows."""
if _slice is None:
_slice = slice(None)
indices = range(*_slice.indices(len(self._rows)))
fields = self.fields
field_index = self._field_index
file_exhausted = False
for i, row in enumerate(self._rows):
# always read next line until EOF to keep in sync
if not file_exhausted:
line: Optional[str] = None
try:
line = next(fh)
except StopIteration:
file_exhausted = True
# now skip if it's not a requested index
if i not in indices:
continue
# proceed only if we have a row in memory or on disk
if row is None:
if line is not None:
row = Row(fields,
tsdb.split(line),
field_index=field_index)
else:
continue
yield (i, row)
def test_split(empty_testsuite):
assert tsdb.split('') == (None, )
assert tsdb.split('one') == ('one', )
assert tsdb.split(u'あ') == (u'あ', )
assert tsdb.split('one@two') == ('one', 'two')
assert tsdb.split('one@@three') == ('one', None, 'three')
assert (tsdb.split('one\\s@\\\\two\\nabc') == ('one@', '\\two\nabc'))
rels = tsdb.read_schema(empty_testsuite)
assert tsdb.split('10@one', fields=rels['item']) == (10, 'one')
def _getitem(self, index: int) -> Row:
"""Get a single non-slice index."""
row = self._rows[index]
if row is None:
# need to handle negative indices manually
if index < 0:
index = len(self._rows) + index
with tsdb.open(self.dir, self.name,
encoding=self.encoding) as lines:
for i, line in enumerate(lines):
if i == index:
row = Row(self.fields,
tsdb.split(line),
field_index=self._field_index)
break
if row is None:
raise ITSDBError('could not retrieve row {}'.format(index))
return row