cassandra.util - Utilities

cassandra.util.datetime_from_timestamp(timestamp)[source]

Creates a timezone-agnostic datetime from timestamp (in seconds) in a consistent manner. Works around a Windows issue with large negative timestamps (PYTHON-119), and rounding differences in Python 3.4 (PYTHON-340).

Parameters:timestamp – a unix timestamp, in seconds
cassandra.util.unix_time_from_uuid1(uuid_arg)[source]

Converts a version 1 uuid.UUID to a timestamp with the same precision as time.time() returns. This is useful for examining the results of queries returning a v1 UUID.

Parameters:uuid_arg – a version 1 UUID
cassandra.util.datetime_from_uuid1(uuid_arg)[source]

Creates a timezone-agnostic datetime from the timestamp in the specified type-1 UUID.

Parameters:uuid_arg – a version 1 UUID
cassandra.util.min_uuid_from_time(timestamp)[source]

Generates the minimum TimeUUID (type 1) for a given timestamp, as compared by Cassandra.

See uuid_from_time() for argument and return types.

cassandra.util.max_uuid_from_time(timestamp)[source]

Generates the maximum TimeUUID (type 1) for a given timestamp, as compared by Cassandra.

See uuid_from_time() for argument and return types.

cassandra.util.uuid_from_time(time_arg, node=None, clock_seq=None)[source]

Converts a datetime or timestamp to a type 1 uuid.UUID.

Parameters:
  • time_arg – The time to use for the timestamp portion of the UUID. This can either be a datetime object or a timestamp in seconds (as returned from time.time()).
  • node (long) – None integer for the UUID (up to 48 bits). If not specified, this field is randomized.
  • clock_seq (int) – Clock sequence field for the UUID (up to 14 bits). If not specified, a random sequence is generated.
Return type:

uuid.UUID

cassandra.util.LOWEST_TIME_UUID = UUID('00000000-0000-1000-8080-808080808080')

The lowest possible TimeUUID, as sorted by Cassandra.

cassandra.util.HIGHEST_TIME_UUID = UUID('ffffffff-ffff-1fff-bf7f-7f7f7f7f7f7f')

The highest possible TimeUUID, as sorted by Cassandra.

class cassandra.util.SortedSet(iterable=())[source]

A sorted set based on sorted list

A sorted set implementation is used in this case because it does not require its elements to be immutable/hashable.

#Not implemented: update functions, inplace operators

cassandra.util.sortedset

alias of SortedSet

class cassandra.util.OrderedMap(*args, **kwargs)[source]

An ordered map that accepts non-hashable types for keys. It also maintains the insertion order of items, behaving as OrderedDict in that regard. These maps are constructed and read just as normal mapping types, exept that they may contain arbitrary collections and other non-hashable items as keys:

>>> od = OrderedMap([({'one': 1, 'two': 2}, 'value'),
...                  ({'three': 3, 'four': 4}, 'value2')])
>>> list(od.keys())
[{'two': 2, 'one': 1}, {'three': 3, 'four': 4}]
>>> list(od.values())
['value', 'value2']

These constructs are needed to support nested collections in Cassandra 2.1.3+, where frozen collections can be specified as parameters to others*:

CREATE TABLE example (
    ...
    value map<frozen<map<int, int>>, double>
    ...
)

This class derives from the (immutable) Mapping API. Objects in these maps are not intended be modified.

* Note: Because of the way Cassandra encodes nested types, when using the driver with nested collections, protocol_version must be 3 or higher.

class cassandra.util.Time(value)[source]

Idealized time, independent of day.

Up to nanosecond resolution

Initializer value can be:

  • integer_type: absolute nanoseconds in the day
  • datetime.time: built-in time
  • string_type: a string time of the form “HH:MM:SS[.mmmuuunnn]”
hour

The hour component of this time (0-23)

minute

The minute component of this time (0-59)

second

The second component of this time (0-59)

nanosecond

The fractional seconds component of the time, in nanoseconds

class cassandra.util.Date(value)[source]

Idealized date: year, month, day

Offers wider year range than datetime.date. For Dates that cannot be represented as a datetime.date (because datetime.MINYEAR, datetime.MAXYEAR), this type falls back to printing days_from_epoch offset.

Initializer value can be:

  • integer_type: absolute days from epoch (1970, 1, 1). Can be negative.
  • datetime.date: built-in date
  • string_type: a string time of the form “yyyy-mm-dd”
seconds

Absolute seconds from epoch (can be negative)

date()[source]

Return a built-in datetime.date for Dates falling in the years [datetime.MINYEAR, datetime.MAXYEAR]

ValueError is raised for Dates outside this range.