API Documentation¶
Streams API¶
The streams API enables you to read data into PyFunctional. The seq function imported with from functional import seq is actually an instance of functional.streams.Stream. Therefore, all the methods available on seq such as seq.csv are documented in the Streams class.

class
functional.streams.
ParallelStream
(processes=None, partition_size=None, disable_compression=False)¶ Bases:
functional.streams.Stream
Parallelized version of functional.streams.Stream normally accessible as pseq

class
functional.streams.
Stream
(disable_compression=False)¶ Bases:
object
Represents and implements a stream which separates the responsibilities of Sequence and ExecutionEngine.
An instance of Stream is normally accessed as seq

csv
(csv_file, dialect='excel', **fmt_params)¶ Reads and parses the input of a csv stream or file.
csv_file can be a filepath or an object that implements the iterator interface (defines next() or __next__() depending on python version).
>>> seq.csv('examples/camping_purchases.csv').take(2) [['1', 'tent', '300'], ['2', 'food', '100']]
Parameters:  csv_file – path to file or iterator object
 dialect – dialect of csv, passed to csv.reader
 fmt_params – options passed to csv.reader
Returns: Sequence wrapping csv file

json
(json_file)¶ Reads and parses the input of a json file handler or file.
Json files are parsed differently depending on if the root is a dictionary or an array.
1) If the json’s root is a dictionary, these are parsed into a sequence of (Key, Value) pairs
2) If the json’s root is an array, these are parsed into a sequence of entries
>>> seq.json('examples/users.json').first() [u'sarah', {u'date_created': u'08/08', u'news_email': True, u'email': u'[email protected]'}]
Parameters: json_file – path or file containing json content Returns: Sequence wrapping jsonl file

jsonl
(jsonl_file)¶ Reads and parses the input of a jsonl file stream or file.
Jsonl formatted files must have a single valid json value on each line which is parsed by the python json module.
>>> seq.jsonl('examples/chat_logs.jsonl').first() {u'date': u'10/09', u'message': u'hello anyone there?', u'user': u'bob'}
Parameters: jsonl_file – path or file containing jsonl content Returns: Sequence wrapping jsonl file

open
(path, delimiter=None, mode='r', buffering=1, encoding=None, errors=None, newline=None)¶ Reads and parses input files as defined.
If delimiter is not None, then the file is read in bulk then split on it. If it is None (the default), then the file is parsed as sequence of lines. The rest of the options are passed directly to builtins.open with the exception that write/append file modes is not allowed.
>>> seq.open('examples/gear_list.txt').take(1) [u'tent\n']
Parameters:  path – path to file
 delimiter – delimiter to split joined text on. if None, defaults to per line split
 mode – file open mode
 buffering – passed to builtins.open
 encoding – passed to builtins.open
 errors – passed to builtins.open
 newline – passed to builtins.open
Returns: output of file depending on options wrapped in a Sequence via seq

range
(*args)¶ Alias to range function where seq.range(args) is equivalent to seq(range(args)).
>>> seq.range(1, 8, 2) [1, 3, 5, 7]
Parameters: args – args to range function Returns: range(args) wrapped by a sequence

sqlite3
(conn, sql, parameters=None, *args, **kwargs)¶ Reads input by querying from a sqlite database.
>>> seq.sqlite3('examples/users.db', 'select id, name from users where id = 1;').first() [(1, 'Tom')]
Parameters:  conn – path or sqlite connection, cursor
 sql – SQL query string
 parameters – Parameters for sql query
Returns: Sequence wrapping SQL cursor

Transformations and Actions API¶
The pipeline module contains the transformations and actions API of PyFunctional

class
functional.pipeline.
Sequence
(sequence, transform=None, engine=None)¶ Bases:
object
Sequence is a wrapper around any type of sequence which provides access to common functional transformations and reductions in a data pipeline style

aggregate
(*args)¶ Aggregates the sequence by specified arguments. Its behavior varies depending on if one, two, or three arguments are passed. Assuming the type of the sequence is A:
One Argument: argument specifies a function of the type f(current: B, next: A => result: B. current represents results computed so far, and next is the next element to aggregate into current in order to return result.
Two Argument: the first argument is the seed value for the aggregation. The second argument is the same as for the one argument case.
Three Argument: the first two arguments are the same as for one and two argument calls. The additional third parameter is a function applied to the result of the aggregation before returning the value.
Parameters: args – options for how to execute the aggregation Returns: aggregated value

all
()¶ Returns True if the truth value of all items in the sequence true.
>>> seq([True, True]).all() True
>>> seq([True, False]).all() False
Returns: True if all items truth value evaluates to True

any
()¶ Returns True if any element in the sequence has truth value True
>>> seq([True, False]).any() True
>>> seq([False, False]).any() False
Returns: True if any element is True

average
(projection=None)¶ Takes the average of elements in the sequence
>>> seq([1, 2]).average() 1.5
>>> seq([('a', 1), ('b', 2)]).average(lambda x: x[1])
Parameters: projection – function to project on the sequence before taking the average Returns: average of elements in the sequence

cache
(delete_lineage=False)¶ Caches the result of the Sequence so far. This means that any functions applied on the pipeline before cache() are evaluated, and the result is stored in the Sequence. This is primarily used internally and is no more helpful than to_list() externally. delete_lineage allows for cache() to be used in internal initialization calls without the caller having knowledge of the internals via the lineage
Parameters: delete_lineage – If set to True, it will cache then erase the lineage

cartesian
(*iterables, **kwargs)¶ Returns the cartesian product of the passed iterables with the specified number of repetitions.
The keyword argument repeat is read from kwargs to pass to itertools.cartesian.
>>> seq.range(2).cartesian(range(2)) [(0, 0), (0, 1), (1, 0), (1, 1)]
Parameters:  iterables – elements for cartesian product
 kwargs – the variable repeat is read from kwargs
Returns: cartesian product

count
(func)¶ Counts the number of elements in the sequence which satisfy the predicate func.
>>> seq([1, 2, 1, 2]).count(lambda x: x > 0) 2
Parameters: func – predicate to count elements on Returns: count of elements that satisfy predicate

dict
(default=None)¶ Converts sequence of (Key, Value) pairs to a dictionary.
>>> type(seq([('a', 1)]).dict()) dict
>>> seq([('a', 1), ('b', 2)]).dict() {'a': 1, 'b': 2}
Parameters: default – Can be a callable zero argument function. When not None, the returned dictionary is a collections.defaultdict with default as value for missing keys. If the value is not callable, then a zero argument lambda function is created returning the value and used for collections.defaultdict Returns: dictionary from sequence of (Key, Value) elements

difference
(other)¶ New sequence with unique elements present in sequence but not in other.
>>> seq([1, 2, 3]).difference([2, 3, 4]) [1]
Parameters: other – sequence to perform difference with Returns: difference of sequence and other

distinct
()¶ Returns sequence of distinct elements. Elements must be hashable.
>>> seq([1, 1, 2, 3, 3, 3, 4]).distinct() [1, 2, 3, 4]
Returns: sequence of distinct elements

distinct_by
(func)¶ Returns sequence of elements who are distinct by the passed function. The return value of func must be hashable. When two elements are distinct by func, the first is taken.
Parameters: func – function to use for determining distinctness Returns: elements distinct by func

drop
(n)¶ Drop the first n elements of the sequence.
>>> seq([1, 2, 3, 4, 5]).drop(2) [3, 4, 5]
Parameters: n – number of elements to drop Returns: sequence without first n elements

drop_right
(n)¶ Drops the last n elements of the sequence.
>>> seq([1, 2, 3, 4, 5]).drop_right(2) [1, 2, 3]
Parameters: n – number of elements to drop Returns: sequence with last n elements dropped

drop_while
(func)¶ Drops elements in the sequence while func evaluates to True, then returns the rest.
>>> seq([1, 2, 3, 4, 5, 1, 2]).drop_while(lambda x: x < 3) [3, 4, 5, 1, 2]
Parameters: func – truth returning function Returns: elements including and after func evaluates to False

empty
()¶ Returns True if the sequence has length zero.
>>> seq([]).empty() True
>>> seq([1]).empty() False
Returns: True if sequence length is zero

enumerate
(start=0)¶ Uses python enumerate to to zip the sequence with indexes starting at start.
>>> seq(['a', 'b', 'c']).enumerate(start=1) [(1, 'a'), (2, 'b'), (3, 'c')]
Parameters: start – Beginning of zip Returns: enumerated sequence starting at start

exists
(func)¶ Returns True if an element in the sequence makes func evaluate to True.
>>> seq([1, 2, 3, 4]).exists(lambda x: x == 2) True
>>> seq([1, 2, 3, 4]).exists(lambda x: x < 0) False
Parameters: func – existence check function Returns: True if any element satisfies func

filter
(func)¶ Filters sequence to include only elements where func is True.
>>> seq([1, 1, 2, 2]).filter(lambda x: x > 0) [1, 2]
Parameters: func – function to filter on Returns: filtered sequence

filter_not
(func)¶ Filters sequence to include only elements where func is False.
>>> seq([1, 1, 2, 2]).filter_not(lambda x: x > 0) [1, 2]
Parameters: func – function to filter_not on Returns: filtered sequence

find
(func)¶ Finds the first element of the sequence that satisfies func. If no such element exists, then return None.
>>> seq(["abc", "ab", "bc"]).find(lambda x: len(x) == 2) 'ab'
Parameters: func – function to find with Returns: first element to satisfy func or None

first
()¶ Returns the first element of the sequence.
>>> seq([1, 2, 3]).first() 1
Raises IndexError when the sequence is empty.
>>> seq([]).first() Traceback (most recent call last): ... IndexError: list index out of range
Returns: first element of sequence

flat_map
(func)¶ Applies func to each element of the sequence, which themselves should be sequences. Then appends each element of each sequence to a final result
>>> seq([[1, 2], [3, 4], [5, 6]]).flat_map(lambda x: x) [1, 2, 3, 4, 5, 6]
>>> seq(["a", "bc", "def"]).flat_map(list) ['a', 'b', 'c', 'd', 'e', 'f']
>>> seq([[1], [2], [3]]).flat_map(lambda x: x * 2) [1, 1, 2, 2, 3, 3]
Parameters: func – function to apply to each sequence in the sequence Returns: application of func to elements followed by flattening

flatten
()¶ Flattens a sequence of sequences to a single sequence of elements.
>>> seq([[1, 2], [3, 4], [5, 6]]) [1, 2, 3, 4, 5, 6]
Returns: flattened sequence

fold_left
(zero_value, func)¶ Assuming that the sequence elements are of type A, folds from left to right starting with the seed value given by zero_value (of type A) using a function of type func(current: B, next: A) => B. current represents the folded value so far and next is the next element from the sequence to fold into current.
>>> seq('a', 'b', 'c').fold_left(['start'], lambda current, next: current + [next])) ['start', 'a', 'b', 'c']
Parameters:  zero_value – zero value to reduce into
 func – Two parameter function as described by function docs
Returns: value from folding values with func into zero_value from left to right.

fold_right
(zero_value, func)¶ Assuming that the sequence elements are of type A, folds from right to left starting with the seed value given by zero_value (of type A) using a function of type func(next: A, current: B) => B. current represents the folded value so far and next is the next element from the sequence to fold into current.
>>> seq('a', 'b', 'c').fold_left(['start'], lambda next, current: current + [next]) ['start', 'c', 'b', a']
Parameters:  zero_value – zero value to reduce into
 func – Two parameter function as described by function docs
Returns: value from folding values with func into zero_value from right to left

for_all
(func)¶ Returns True if all elements in sequence make func evaluate to True.
>>> seq([1, 2, 3]).for_all(lambda x: x > 0) True
>>> seq([1, 2, 1]).for_all(lambda x: x > 0) False
Parameters: func – function to check truth value of all elements with Returns: True if all elements make func evaluate to True

for_each
(func)¶ Executes func on each element of the sequence.
>>> l = [] >>> seq([1, 2, 3, 4]).for_each(l.append) >>> l [1, 2, 3, 4]
Parameters: func – function to execute

group_by
(func)¶ Group elements into a list of (Key, Value) tuples where func creates the key and maps to values matching that key.
>>> seq(["abc", "ab", "z", "f", "qw"]).group_by(len) [(1, ['z', 'f']), (2, ['ab', 'qw']), (3, ['abc'])]
Parameters: func – group by result of this function Returns: grouped sequence

group_by_key
()¶ Group sequence of (Key, Value) elements by Key.
>>> seq([('a', 1), ('b', 2), ('b', 3), ('b', 4), ('c', 3), ('c', 0)]).group_by_key() [('a', [1]), ('c', [3, 0]), ('b', [2, 3, 4])]
Returns: sequence grouped by key

grouped
(size)¶ Partitions the elements into groups of length size.
>>> seq([1, 2, 3, 4, 5, 6, 7, 8]).grouped(2) [[1, 2], [3, 4], [5, 6], [7, 8]]
>>> seq([1, 2, 3, 4, 5, 6, 7, 8]).grouped(3) [[1, 2, 3], [4, 5, 6], [7, 8]]
The last partition has at least one element but may have less than size elements.
Parameters: size – size of the partitions Returns: sequence partitioned into groups of length size

head
()¶ Returns the first element of the sequence.
>>> seq([1, 2, 3]).head() 1
Raises IndexError when the sequence is empty.
>>> seq([]).head() Traceback (most recent call last): ... IndexError: list index out of range
Returns: first element of sequence

head_option
()¶ Returns the first element of the sequence or None, if the sequence is empty.
>>> seq([1, 2, 3]).head_option() 1
>>> seq([]).head_option() None
Returns: first element of sequence or None if sequence is empty

init
()¶ Returns the sequence, without its last element.
>>> seq([1, 2, 3]).init() [1, 2]
Returns: sequence without last element

inits
()¶ Returns consecutive inits of the sequence.
>>> seq([1, 2, 3]).inits() [[1, 2, 3], [1, 2], [1], []]
Returns: consecutive init()s on sequence

inner_join
(other)¶ Sequence and other must be composed of (Key, Value) pairs. If self.sequence contains (K, V) pairs and other contains (K, W) pairs, the return result is a sequence of (K, (V, W)) pairs. Will return only elements where the key exists in both sequences.
>>> seq([('a', 1), ('b', 2), ('c', 3)]).inner_join([('a', 2), ('c', 5)]) [('a', (1, 2)), ('c', (3, 5))]
Parameters: other – sequence to join with Returns: joined sequence of (K, (V, W)) pairs

intersection
(other)¶ New sequence with unique elements present in sequence and other.
>>> seq([1, 1, 2, 3]).intersection([2, 3, 4]) [2, 3]
Parameters: other – sequence to perform intersection with Returns: intersection of sequence and other

join
(other, join_type='inner')¶ Sequence and other must be composed of (Key, Value) pairs. If self.sequence contains (K, V) pairs and other contains (K, W) pairs, the return result is a sequence of (K, (V, W)) pairs. If join_type is “left”, V values will always be present, W values may be present or None. If join_type is “right”, W values will always be present, W values may be present or None. If join_type is “outer”, V or W may be present or None, but never at the same time.
>>> seq([('a', 1), ('b', 2), ('c', 3)]).join([('a', 2), ('c', 5)], "inner") [('a', (1, 2)), ('c', (3, 5))]
>>> seq([('a', 1), ('b', 2), ('c', 3)]).join([('a', 2), ('c', 5)]) [('a', (1, 2)), ('c', (3, 5))]
>>> seq([('a', 1), ('b', 2)]).join([('a', 3), ('c', 4)], "left") [('a', (1, 3)), ('b', (2, None)]
>>> seq([('a', 1), ('b', 2)]).join([('a', 3), ('c', 4)], "right") [('a', (1, 3)), ('c', (None, 4)]
>>> seq([('a', 1), ('b', 2)]).join([('a', 3), ('c', 4)], "outer") [('a', (1, 3)), ('b', (2, None)), ('c', (None, 4))]
Parameters:  other – sequence to join with
 join_type – specifies join_type, may be “left”, “right”, or “outer”
Returns: side joined sequence of (K, (V, W)) pairs

last
()¶ Returns the last element of the sequence.
>>> seq([1, 2, 3]).last() 3
Raises IndexError when the sequence is empty.
>>> seq([]).last() Traceback (most recent call last): ... IndexError: list index out of range
Returns: last element of sequence

last_option
()¶ Returns the last element of the sequence or None, if the sequence is empty.
>>> seq([1, 2, 3]).last_option() 3
>>> seq([]).last_option() None
Returns: last element of sequence or None if sequence is empty

left_join
(other)¶ Sequence and other must be composed of (Key, Value) pairs. If self.sequence contains (K, V) pairs and other contains (K, W) pairs, the return result is a sequence of (K, (V, W)) pairs. V values will always be present, W values may be present or None.
>>> seq([('a', 1), ('b', 2)]).join([('a', 3), ('c', 4)]) [('a', (1, 3)), ('b', (2, None)]
Parameters: other – sequence to join with Returns: left joined sequence of (K, (V, W)) pairs

len
()¶ Return length of sequence using its length function.
>>> seq([1, 2, 3]).len() 3
Returns: length of sequence

list
()¶ Converts sequence to list of elements.
>>> type(seq([]).list()) list
>>> type(seq([])) functional.pipeline.Sequence
>>> seq([1, 2, 3]).list() [1, 2, 3]
Returns: list of elements in sequence

make_string
(separator)¶ Concatenate the elements of the sequence into a string separated by separator.
>>> seq([1, 2, 3]).make_string("@") '[email protected]@3'
Parameters: separator – string separating elements in string Returns: concatenated string separated by separator

map
(func)¶ Maps f onto the elements of the sequence.
>>> seq([1, 2, 3, 4]).map(lambda x: x * 1) [1, 2, 3, 4]
Parameters: func – function to map with Returns: sequence with func mapped onto it

max
()¶ Returns the largest element in the sequence. If the sequence has multiple maximal elements, only the first one is returned.
The compared objects must have defined comparison methods. Raises TypeError when the objects are not comparable.
The sequence can not be empty. Raises ValueError when the sequence is empty.
>>> seq([2, 4, 5, 1, 3]).max() 5
>>> seq('aa', 'xyz', 'abcd', 'xyy').max() 'xyz'
>>> seq([1, "a"]).max() Traceback (most recent call last): ... TypeError: unorderable types: int() < str()
>>> seq([]).max() Traceback (most recent call last): ... ValueError: max() arg is an empty sequence
Returns: Maximal value of sequence

max_by
(func)¶ Returns the largest element in the sequence. Provided function is used to generate key used to compare the elements. If the sequence has multiple maximal elements, only the first one is returned.
The sequence can not be empty. Raises ValueError when the sequence is empty.
>>> seq([2, 4, 5, 1, 3]).max_by(lambda num: num % 4) 3
>>> seq('aa', 'xyz', 'abcd', 'xyy').max_by(len) 'abcd'
>>> seq([]).max_by(lambda x: x) Traceback (most recent call last): ... ValueError: max() arg is an empty sequence
Parameters: func – function to compute max by Returns: Maximal element by func(element)

min
()¶ Returns the smallest element in the sequence. If the sequence has multiple minimal elements, only the first one is returned.
The compared objects must have defined comparison methods. Raises TypeError when the objects are not comparable.
The sequence can not be empty. Raises ValueError when the sequence is empty.
>>> seq([2, 4, 5, 1, 3]).min() 1
>>> seq('aa', 'xyz', 'abcd', 'xyy').min() 'aa'
>>> seq([1, "a"]).min() Traceback (most recent call last): ... TypeError: unorderable types: int() < str()
>>> seq([]).min() Traceback (most recent call last): ... ValueError: min() arg is an empty sequence
Returns: Minimal value of sequence

min_by
(func)¶ Returns the smallest element in the sequence. Provided function is used to generate key used to compare the elements. If the sequence has multiple minimal elements, only the first one is returned.
The sequence can not be empty. Raises ValueError when the sequence is empty.
>>> seq([2, 4, 5, 1, 3]).min_by(lambda num: num % 6) 5
>>> seq('aa', 'xyz', 'abcd', 'xyy').min_by(len) 'aa'
>>> seq([]).min_by(lambda x: x) Traceback (most recent call last): ... ValueError: min() arg is an empty sequence
Parameters: func – function to compute min by Returns: Maximal element by func(element)

non_empty
()¶ Returns True if the sequence does not have length zero.
>>> seq([]).non_empty() False
>>> seq([1]).non_empty() True
Returns: True if sequence length is not zero

order_by
(func)¶ Orders the input according to func
>>> seq([(2, 'a'), (1, 'b'), (4, 'c'), (3, 'd')]).order_by(lambda x: x[0]) [1, 2, 3, 4]
Parameters: func – order by funciton Returns: ordered sequence

outer_join
(other)¶ Sequence and other must be composed of (Key, Value) pairs. If self.sequence contains (K, V) pairs and other contains (K, W) pairs, the return result is a sequence of (K, (V, W)) pairs. One of V or W will always be not None, but the other may be None
>>> seq([('a', 1), ('b', 2)]).outer_join([('a', 3), ('c', 4)], "outer") [('a', (1, 3)), ('b', (2, None)), ('c', (None, 4))]
Parameters: other – sequence to join with Returns: outer joined sequence of (K, (V, W)) pairs

partition
(func)¶ Partition the sequence based on satisfying the predicate func.
>>> seq([1, 1, 2, 2]).partition(lambda x: x < 0) ([1, 2], [1, 2])
Parameters: func – predicate to partition on Returns: tuple of partitioned sequences

product
(projection=None)¶ Takes product of elements in sequence.
>>> seq([1, 2, 3, 4]).product() 24
>>> seq([]).product() 1
>>> seq([(1, 2), (1, 3), (1, 4)]).product(lambda x: x[0]) 1
Parameters: projection – function to project on the sequence before taking the product Returns: product of elements in sequence

reduce
(func)¶ Reduce sequence of elements using func.
>>> seq([1, 2, 3]).reduce(lambda x, y: x + y) 6
Parameters: func – two parameter, associative reduce function Returns: reduced value using func

reduce_by_key
(func)¶ Reduces a sequence of (Key, Value) using func on each sequence of values.
>>> seq([('a', 1), ('b', 2), ('b', 3), ('b', 4), ('c', 3), ('c', 0)]) .reduce_by_key(lambda x, y: x + y) [('a', 1), ('c', 3), ('b', 9)]
Parameters: func – reduce each list of values using two parameter, associative func Returns: Sequence of tuples where the value is reduced with func

reverse
()¶ Returns the reversed sequence.
>>> seq([1, 2, 3]).reverse() [3, 2, 1]
Returns: reversed sequence

right_join
(other)¶ Sequence and other must be composed of (Key, Value) pairs. If self.sequence contains (K, V) pairs and other contains (K, W) pairs, the return result is a sequence of (K, (V, W)) pairs. W values will always bepresent, V values may be present or None.
>>> seq([('a', 1), ('b', 2)]).join([('a', 3), ('c', 4)]) [('a', (1, 3)), ('b', (2, None)]
Parameters: other – sequence to join with Returns: right joined sequence of (K, (V, W)) pairs

select
(func)¶ Selects f from the elements of the sequence.
>>> seq([1, 2, 3, 4]).select(lambda x: x * 1) [1, 2, 3, 4]
Parameters: func – function to select with Returns: sequence with func mapped onto it

sequence
¶ Alias for to_list used internally for brevity
Returns: result of to_list() on sequence

set
()¶ Converts sequence to a set of elements.
>>> type(seq([])).to_set() set
>>> type(seq([])) functional.pipeline.Sequence
>>> seq([1, 1, 2, 2]).set() {1, 2}
:return:set of elements in sequence

size
()¶ Return size of sequence using its length function.
Returns: size of sequence

slice
(start, until)¶ Takes a slice of the sequence starting at start and until but not including until.
>>> seq([1, 2, 3, 4]).slice(1, 2) [2] >>> seq([1, 2, 3, 4]).slice(1, 3) [2, 3]
Parameters:  start – starting index
 until – ending index
Returns: slice including start until but not including until

sliding
(size, step=1)¶ Groups elements in fixed size blocks by passing a sliding window over them.
The last window has at least one element but may have less than size elements
Parameters:  size – size of sliding window
 step – step size between windows
Returns: sequence of sliding windows

sorted
(key=None, reverse=False)¶ Uses python sort and its passed arguments to sort the input.
>>> seq([2, 1, 4, 3]).sorted() [1, 2, 3, 4]
Parameters:  key – sort using key function
 reverse – return list reversed or not
Returns: sorted sequence

sum
(projection=None)¶ Takes sum of elements in sequence.
>>> seq([1, 2, 3, 4]).sum() 10
>>> seq([(1, 2), (1, 3), (1, 4)]).sum(lambda x: x[0]) 3
Parameters: projection – function to project on the sequence before taking the sum Returns: sum of elements in sequence

symmetric_difference
(other)¶ New sequence with elements in either sequence or other, but not both.
>>> seq([1, 2, 3, 3]).symmetric_difference([2, 4, 5]) [1, 3, 4, 5]
Parameters: other – sequence to perform symmetric difference with Returns: symmetric difference of sequence and other

tail
()¶ Returns the sequence, without its first element.
>>> seq([1, 2, 3]).init() [2, 3]
Returns: sequence without first element

tails
()¶ Returns consecutive tails of the sequence.
>>> seq([1, 2, 3]).tails() [[1, 2, 3], [2, 3], [3], []]
Returns: consecutive tail()s of the sequence

take
(n)¶ Take the first n elements of the sequence.
>>> seq([1, 2, 3, 4]).take(2) [1, 2]
Parameters: n – number of elements to take Returns: first n elements of sequence

take_while
(func)¶ Take elements in the sequence until func evaluates to False, then return them.
>>> seq([1, 2, 3, 4, 5, 1, 2]).take_while(lambda x: x < 3) [1, 2]
Parameters: func – truth returning function Returns: elements taken until func evaluates to False

to_csv
(path, mode='wb', dialect='excel', compression=None, **fmtparams)¶ Saves the sequence to a csv file. Each element should be an iterable which will be expanded to the elements of each row.
Parameters:  path – path to write file
 mode – file open mode
 dialect – passed to csv.writer
 fmtparams – passed to csv.writer

to_dict
(default=None)¶ Converts sequence of (Key, Value) pairs to a dictionary.
>>> type(seq([('a', 1)]).to_dict()) dict
>>> seq([('a', 1), ('b', 2)]).to_dict() {'a': 1, 'b': 2}
Parameters: default – Can be a callable zero argument function. When not None, the returned dictionary is a collections.defaultdict with default as value for missing keys. If the value is not callable, then a zero argument lambda function is created returning the value and used for collections.defaultdict Returns: dictionary from sequence of (Key, Value) elements

to_file
(path, delimiter=None, mode='wt', buffering=1, encoding=None, errors=None, newline=None, compresslevel=9, format=None, check=1, preset=None, filters=None, compression=None)¶ Saves the sequence to a file by executing str(self) which becomes str(self.to_list()). If delimiter is defined will instead execute self.make_string(delimiter)
Parameters:  path – path to write file
 delimiter – if defined, will call make_string(delimiter) and save that to file.
 mode – file open mode
 buffering – passed to builtins.open
 encoding – passed to builtins.open
 errors – passed to builtins.open
 newline – passed to builtins.open
 compression – compression format
 compresslevel – passed to gzip.open
 format – passed to lzma.open
 check – passed to lzma.open
 preset – passed to lzma.open
 filters – passed to lzma.open

to_json
(path, root_array=True, mode='wb', compression=None)¶ Saves the sequence to a json file. If root_array is True, then the sequence will be written to json with an array at the root. If it is False, then the sequence will be converted from a sequence of (Key, Value) pairs to a dictionary so that the json root is a dictionary.
Parameters:  path – path to write file
 root_array – write json root as an array or dictionary
 mode – file open mode

to_jsonl
(path, mode='wb', compression=None)¶ Saves the sequence to a jsonl file. Each element is mapped using json.dumps then written with a newline separating each element.
Parameters:  path – path to write file
 mode – mode to write in, defaults to ‘w’ to overwrite contents
 compression – compression format

to_list
()¶ Converts sequence to list of elements.
>>> type(seq([]).to_list()) list
>>> type(seq([])) functional.pipeline.Sequence
>>> seq([1, 2, 3]).to_list() [1, 2, 3]
Returns: list of elements in sequence

to_pandas
(columns=None)¶ Converts sequence to a pandas DataFrame using pandas.DataFrame.from_records
Parameters: columns – columns for pandas to use Returns: DataFrame of sequence

to_set
()¶ Converts sequence to a set of elements.
>>> type(seq([])).to_set() set
>>> type(seq([])) functional.pipeline.Sequence
>>> seq([1, 1, 2, 2]).to_set() {1, 2}
:return:set of elements in sequence

to_sqlite3
(conn, target, *args, **kwargs)¶ Saves the sequence to sqlite3 database. Target table must be created in advance. The table schema is inferred from the elements in the sequence if only target table name is supplied.
>>> seq([(1, 'Tom'), (2, 'Jack')]) .to_sqlite3('users.db', 'INSERT INTO user (id, name) VALUES (?, ?)')
>>> seq([{'id': 1, 'name': 'Tom'}, {'id': 2, 'name': 'Jack'}]).to_sqlite3(conn, 'user')
Parameters:  conn – path or sqlite connection, cursor
 target – SQL query string or table name
 args – passed to sqlite3.connect
 kwargs – passed to sqlite3.connect

union
(other)¶ New sequence with unique elements from self and other.
>>> seq([1, 1, 2, 3, 3]).union([1, 4, 5]) [1, 2, 3, 4, 5]
Parameters: other – sequence to union with Returns: union of sequence and other

where
(func)¶ Selects elements where func evaluates to True.
>>> seq([1, 1, 2, 2]).where(lambda x: x > 0) [1, 2]
Parameters: func – function to filter on Returns: filtered sequence

zip
(sequence)¶ Zips the stored sequence with the given sequence.
>>> seq([1, 2, 3]).zip([4, 5, 6]) [(1, 4), (2, 5), (3, 6)]
Parameters: sequence – second sequence to zip Returns: stored sequence zipped with given sequence

zip_with_index
(start=0)¶ Zips the sequence to its index, with the index being the second element of each tuple.
>>> seq(['a', 'b', 'c']).zip_with_index() [('a', 0), ('b', 1), ('c', 2)]
Returns: sequence zipped to its index
