Operators

Nextflow operators are methods that allow you to manipulate channels. Every operator, with the exception of set and subscribe, produces one or more new channels, allowing you to chain operators to fit your needs.

This page is a comprehensive reference for all Nextflow operators. However, if you are new to Nextflow, here are some suggested operators to learn for common use cases:

Filtering: filter, randomSample, take, unique
Reduction: collect, groupTuple, reduce
Text processing: splitCsv, splitJson, splitText
Combining channels: combine, concat, join, mix
Forking channels: branch, multiMap
Maths: count, max, min, sum
Other: ifEmpty, map, set, view

branch

New in version 19.08.0-edge.

Returns: map of queue channels

The branch operator forwards each item from a source channel to one of multiple output channels, based on a selection criteria.

The selection criteria is a closure that defines, for each output channel, a unique label followed by a boolean expression. When an item is received, it is routed to the first output channel whose expression evaluates to true. For example:

Channel.of(1, 2, 3, 40, 50)
    .branch {
        small: it < 10
        large: it > 10
    }
    .set { result }

result.small.view { "$it is small" }
result.large.view { "$it is large" }

is small
is small
is small
is large
is large

Note

The above output may be printed in any order since the two view operations are executed asynchronously.

A fallback condition can be specified using true as the last branch condition:

Channel.of(1, 2, 3, 40, 50)
    .branch {
        small: it < 10
        large: it < 50
        other: true
    }
    .set { result }

result.small.view { "$it is small" }
result.large.view { "$it is large" }
result.other.view { "$it is other" }

is small
is small
is small
is large
is other

The value emitted to each branch can be customized with an expression statement (or statements) after the branch condition:

Channel.of(1, 2, 3, 40, 50)
    .branch {
        foo: it < 10
            return it+2

        bar: it < 50
            return it-2

        other: true
            return 0
    }
    .set { result }

result.foo.view { "$it is foo" }
result.bar.view { "$it is bar" }
result.other.view { "$it is other" }

is foo
is foo
is foo
is bar
is other

Tip

When the return keyword is omitted, the value of the last expression statement is implicitly returned.

The branchCriteria() method can be used to create a branch criteria as a variable that can be passed as an argument to any number of branch operations, as shown below:

def criteria = branchCriteria {
    small: it < 10
    large: it > 10
}

Channel.of(1, 2, 30).branch(criteria).set { ch1 }
Channel.of(10, 20, 3).branch(criteria).set { ch2 }

ch1.small.view { "$it is small" }
ch1.large.view { "$it is large" }
ch2.small.view { "$it is small" }
ch2.large.view { "$it is large" }

is small
is small
is small
is large
is large

buffer

Returns: queue channel

The buffer operator collects items from a source channel into subsets and emits each subset separately.

This operator has multiple variants:

buffer( closingCondition )

Emits each subset when closingCondition is satisfied. The closing condition can be a literal value, a regular expression, a type qualifier (i.e. Java class), or a boolean predicate. For example:

Channel.of( 1, 2, 3, 1, 2, 3 )
    .buffer { it == 2 }
    .view()

[1, 2]
[3, 1, 2]

buffer( openingCondition, closingCondition )

Creates a new subset when openingCondition is satisfied and emits the subset when is closingCondition is satisfied. The opening and closing conditions can each be a literal value, a regular expression, a type qualifier (i.e. Java class), or a boolean predicate. For example:

// emits bundles starting with `2` and ending with `4`
Channel.of( 1, 2, 3, 4, 5, 1, 2, 3, 4, 5, 1, 2 )
    .buffer( 2, 4 )
    .view()

[2, 3, 4]
[2, 3, 4]

buffer( size: n, remainder: true | false )

Emits a new subset for every n items. Remaining items are discarded. For example:

Channel.of( 1, 2, 3, 1, 2, 3, 1 )
    .buffer( size: 2 )
    .view()

[1, 2]
[3, 1]
[2, 3]

The remainder option can be used to emit any remaining items as a partial subset:

Channel.of( 1, 2, 3, 1, 2, 3, 1 )
    .buffer( size: 2, remainder: true )
    .view()

[1, 2]
[3, 1]
[2, 3]
[1]

buffer( size: n, skip: m, remainder: true | false )

Emits a new subset for every n items, skipping m items before collecting each subset. For example:

Channel.of( 1, 2, 3, 4, 5, 1, 2, 3, 4, 5, 1, 2 )
    .buffer( size: 3, skip: 2 )
    .view()

[3, 4, 5]
[3, 4, 5]

The remainder option can be used to emit any remaining items as a partial subset.

collate

Returns: queue channel

The collate operator collects items from a source channel into groups of N items.

This operator has multiple variants:

collate( size, remainder = true )

Collects items into groups of size items:

Channel.of(1, 2, 3, 1, 2, 3, 1)
    .collate( 3 )
    .view()

[1, 2, 3]
[1, 2, 3]
[1]

By default, any remaining items are emitted as a partial group. You can specify false as the second parameter to discard them instead:

Channel.of(1, 2, 3, 1, 2, 3, 1)
    .collate( 3, false )
    .view()

[1, 2, 3]
[1, 2, 3]

Note

This version of collate is equivalent to buffer( size: n, remainder: true | false ).

collate( size, step, remainder = true )

Collects items into groups of size items using a sliding window that moves by step items at a time:

Channel.of(1, 2, 3, 4)
    .collate( 3, 1 )
    .view()

[1, 2, 3]
[2, 3, 4]
[3, 4]
[4]

You can specify false as the third parameter to discard any remaining items.

collect

Returns: value channel

The collect operator collects all items from a source channel into a list and emits it as a single item:

Channel.of( 1, 2, 3, 4 )
    .collect()
    .view()

[1, 2, 3, 4]

An optional closure can be used to transform each item before it is collected:

Channel.of( 'hello', 'ciao', 'bonjour' )
    .collect { it.length() }
    .view()

[5, 4, 7]

Available options:

flat: When true, nested list structures are flattened and their items are collected individually (default: true).
sort: When true, the collected items are sorted by their natural ordering (default: false). Can also be a closure or a Comparator which defines how items are compared during sorting.

collectFile

Returns: queue channel

The collectFile operator collects the items from a source channel and saves them to one or more files, emitting the collected file(s).

This operator has multiple variants:

collectFile( name: '...', options = [:] )

Collects the items and saves them to a single file specified by the name option:

Channel.of('alpha', 'beta', 'gamma')
    .collectFile(name: 'sample.txt', newLine: true)
    .subscribe {
        println "Entries are saved to file: $it"
        println "File content is: ${it.text}"
    }

collectFile( closure, options = [:] )

Collects the items into groups and saves each group to a file, using a grouping criteria. The grouping criteria is a closure that maps each item to a pair, where the first element is the file name for the group and the second element is the content to be appended to that file. For example:

Channel.of('Hola', 'Ciao', 'Hello', 'Bonjour', 'Halo')
    .collectFile { item ->
        [ "${item[0]}.txt", item + '\n' ]
    }
    .subscribe {
        println "File '${it.name}' contains:"
        println it.text
    }

File 'B.txt' contains:
Bonjour

File 'C.txt' contains:
Ciao

File 'H.txt' contains:
Halo
Hola
Hello

When the items from the source channel are files, the grouping criteria can be omitted. In this case, the items will be grouped by their source filename.

The following example shows how to use a closure to collect and sort all sequences in a FASTA file from shortest to longest:

Channel
    .fromPath('/data/sequences.fa')
    .splitFasta( record: [id: true, sequence: true] )
    .collectFile( name: 'result.fa', sort: { it.size() } ) {
        it.sequence
    }
    .view { it.text }

Warning

The collectFile operator needs to store files in a temporary directory that is automatically deleted on workflow completion. For performance reasons, this directory is located in the machine’s local storage, and it should have as much free space as the data that is being collected. The tempDir option can be used to specify a different temporary directory.

Available options:

cache

Controls the caching ability of the collectFile operator when using the resume feature. It follows the same semantic of the cache directive (default: true).

keepHeader

Prepend the resulting file with the header fetched in the first collected file. The header size (ie. lines) can be specified by using the skip option (default: 0), to determine how many lines to remove from all collected files except for the first (where no lines will be removed).

name

Name of the file where all received values are stored.

newLine

Appends a newline character automatically after each entry (default: false).

seed

A value or a map of values used to initialize the files content.

skip

Skip the first n lines e.g. skip: 1 (default: 0).

sort

Defines sorting criteria of content in resulting file(s). Can be one of the following values:

false: Disable content sorting. Entries are appended as they are produced.
true: Order the content by the entry’s natural ordering i.e. numerical for number, lexicographic for string, etc. See the Java documentation for more information.
'index': Order the content by the incremental index number assigned to each entry while they are collected.
'hash': (default) Order the content by the hash number associated to each entry
'deep': Similar to the previous, but the hash number is created on actual entries content e.g. when the entry is a file the hash is created on the actual file content.
A custom sorting criteria can be specified with a Closure or a Comparator object.

The file content is sorted in such a way that it does not depend on the order in which entries were added to it, which guarantees that it is consistent (i.e. does not change) across different executions with the same data.

storeDir

Folder where the resulting file(s) are stored.

tempDir

Folder where temporary files, used by the collecting process, are stored.

combine

Returns: queue channel

The combine operator produces the combinations (i.e. cross product, “Cartesian” product) of two source channels, or a channel and a list (as the right operand), emitting each combination separately.

For example:

numbers = Channel.of(1, 2, 3)
words = Channel.of('hello', 'ciao')

numbers
    .combine(words)
    .view()

[1, hello]
[2, hello]
[3, hello]
[1, ciao]
[2, ciao]
[3, ciao]

The by option can be used to combine items that share a matching key. The value should be the zero-based index of the tuple, or a list of indices. For example:

source = Channel.of( [1, 'alpha'], [2, 'beta'] )
target = Channel.of( [1, 'x'], [1, 'y'], [1, 'z'], [2, 'p'], [2, 'q'], [2, 't'] )

source.combine(target, by: 0).view()

[1, alpha, x]
[1, alpha, y]
[1, alpha, z]
[2, beta, p]
[2, beta, q]
[2, beta, t]

Note

The combine operator is similar to cross and join, making them easy to confuse. Their differences can be summarized as follows:

combine and cross both produce an outer product or cross product, whereas join produces an inner product.
combine filters pairs with a matching key only if the by option is used, whereas cross always filters pairs with a matching key.
combine with the by option merges and flattens each pair, whereas cross does not. Compare the examples for combine and cross to see this difference.

concat

Returns: queue channel

The concat operator emits the items from two or more source channels into a single output channel. Each source channel is emitted in the order in which it was specified.

In other words, given N channels, the items from the i+1-th channel are emitted only after all of the items from the i-th channel have been emitted.

For example:

a = Channel.of( 'a', 'b', 'c' )
b = Channel.of( 1, 2, 3 )
c = Channel.of( 'p', 'q' )

c.concat( b, a ).view()

p
q
1
2
3
a
b
c

count

Returns: value channel

The count operator computes the total number of items in a source channel and emits it:

Channel.of(9, 1, 7, 5)
    .count()
    .view()

An optional filter can be provided to select which items to count. The selection criteria can be a literal value, a regular expression, a type qualifier (i.e. Java class), or a boolean predicate. For example:

Channel.of(4, 1, 7, 1, 1)
    .count(1)
    .view()

Channel.of('a', 'c', 'c', 'q', 'b')
    .count( ~/c/ )
    .view()

Channel.of('a', 'c', 'c', 'q', 'b')
    .count { it <= 'c' }
    .view()

countFasta

Returns: value channel

Counts the total number of records in a channel of FASTA files, equivalent to splitFasta | count. See splitFasta for the list of available options.

countFastq

Returns: value channel

Counts the total number of records in a channel of FASTQ files, equivalent to splitFastq | count. See splitFastq for the list of available options.

countJson

Returns: value channel

Counts the total number of records in a channel of JSON files, equivalent to splitJson | count. See splitJson for the list of available options.

countLines

Returns: value channel

Counts the total number of lines in a channel of text files, equivalent to splitText | count. See splitLines for the list of available options.

cross

Returns: queue channel

The cross operator emits every pairwise combination of two channels for which the pair has a matching key.

By default, the key is defined as the first entry in a list or map, or the value itself for any other data type. For example:

source = Channel.of( [1, 'alpha'], [2, 'beta'] )
target = Channel.of( [1, 'x'], [1, 'y'], [1, 'z'], [2, 'p'], [2, 'q'], [2, 't'] )

source.cross(target).view()

[[1, alpha], [1, x]]
[[1, alpha], [1, y]]
[[1, alpha], [1, z]]
[[2, beta], [2, p]]
[[2, beta], [2, q]]
[[2, beta], [2, t]]

An optional closure can be used to define the matching key for each item:

source = Channel.of( [1, 'alpha'], [2, 'beta'] )
target = Channel.of( [1, 'a'], [1, 'b'], [2, 'a'], [2, 'b'] )

source .cross(target) { it[1][0] } .view()

[[1, alpha], [1, a]]
[[1, alpha], [2, a]]
[[2, beta], [1, b]]
[[2, beta], [2, b]]

There are two important caveats when using the cross operator:

The operator is not commutative, i.e. a.cross(b) is not the same as b.cross(a)
Each source channel should not emit any items with duplicate keys, i.e. each item should have a unique key.

distinct

Returns: queue channel

The distinct operator forwards a source channel with consecutively repeated items removed, such that each emitted item is different from the preceding one:

Channel.of( 1, 1, 2, 2, 2, 3, 1, 1, 2, 2, 3 )
    .distinct()
    .view()

An optional closure can be used to transform each value before it is evaluated for distinct-ness:

Channel.of( 1, 1, 2, 2, 2, 3, 1, 1, 2, 4, 6 )
    .distinct { it % 2 }
    .view()

dump

Returns: queue channel or value channel, depending on the input

The dump operator prints each item in a source channel when the pipeline is executed with the -dump-channels command-line option, otherwise it does nothing. It is a useful way to inspect and debug channels quickly without having to modify the pipeline script.

The tag option can be used to select which channels to dump:

Channel.of( 1, 2, 3 )
    .map { it+1 }
    .dump(tag: 'foo')

Channel.of( 1, 2, 3 )
    .map { it^2 }
    .dump(tag: 'bar')

Then, you can run your pipeline with -dump-channels foo or -dump-channels bar to dump the content of either channel. Multiple tag names can be specified as a comma-separated list.

Available options:

pretty: New in version 22.10.0.; When true, format the output as pretty-printed JSON (default: false).
tag: Associate the channel with a tag that can be specified with the -dump-channels option to select which channels to dump.

filter

Returns: queue channel

The filter operator emits the items from a source channel that satisfy a condition, discarding all other items. The filter condition can be a literal value, a regular expression, a type qualifier (i.e. Java class), or a boolean predicate.

The following example filters a channel with a regular expression that only matches strings beginning with a:

Channel.of( 'a', 'b', 'aa', 'bc', 3, 4.5 )
    .filter( ~/^a.*/ )
    .view()

a
aa

The following example filters a channel with the Number type qualifier so that only numbers are emitted:

Channel.of( 'a', 'b', 'aa', 'bc', 3, 4.5 )
    .filter( Number )
    .view()

3
4.5

The following example filters a channel using a boolean predicate, which is a closure that returns a boolean value. In this case, the predicate is used to select only odd numbers:

Channel.of( 1, 2, 3, 4, 5 )
    .filter { it % 2 == 1 }
    .view()

1
3
5

first

Returns: value channel

The first operator emits the first item in a source channel, or the first item that matches a condition. The condition can be a regular expression, a type qualifier (i.e. Java class), or a boolean predicate. For example:

// no condition is specified, emits the very first item: 1
Channel.of( 1, 2, 3 )
    .first()
    .view()

// emits the first item matching the regular expression: 'aa'
Channel.of( 'a', 'aa', 'aaa' )
    .first( ~/aa.*/ )
    .view()

// emits the first String value: 'a'
Channel.of( 1, 2, 'a', 'b', 3 )
    .first( String )
    .view()

// emits the first item for which the predicate evaluates to true: 4
Channel.of( 1, 2, 3, 4, 5 )
    .first { it > 3 }
    .view()

flatMap

Returns: queue channel

The flatMap operator applies a mapping function to each item from a source channel.

When the mapping function returns a list, each element in the list is emitted separately:

Channel.of( 1, 2, 3 )
    .flatMap { n -> [ n, n*2, n*3 ] }
    .view()

When the mapping function returns a map, each key-value pair in the map is emitted separately:

Channel.of( 1, 2, 3 )
    .flatMap { n -> [ number: n, square: n*n, cube: n*n*n ] }
    .view { "${it.key}: ${it.value}" }

number: 1
square: 1
cube: 1
number: 2
square: 4
cube: 8
number: 3
square: 9
cube: 27

flatten

Returns: queue channel

The flatten operator flattens each item from a source channel that is a list or other collection, such that each element in each collection is emitted separately:

Channel.of( [1, [2, 3]], 4, [5, [6]] )
    .flatten()
    .view()

As shown in the above example, deeply nested collections are also flattened.

groupTuple

Returns: queue channel

The groupTuple operator collects lists (i.e. tuples) from a source channel into groups based on a grouping key. A new tuple is emitted for each distinct key.

To be more precise, the operator transforms a sequence of tuples like (K, V, W, ..) into a sequence of tuples like (K, list(V), list(W), ..).

For example:

Channel.of( [1, 'A'], [1, 'B'], [2, 'C'], [3, 'B'], [1, 'C'], [2, 'A'], [3, 'D'] )
    .groupTuple()
    .view()

[1, [A, B, C]]
[2, [C, A]]
[3, [B, D]]

By default, the first element of each tuple is used as the grouping key. The by option can be used to specify a different index, or list of indices. For example, to group by the second element of each tuple:

Channel.of( [1, 'A'], [1, 'B'], [2, 'C'], [3, 'B'], [1, 'C'], [2, 'A'], [3, 'D'] )
    .groupTuple(by: 1)
    .view()

[[1, 2], A]
[[1, 3], B]
[[2, 1], C]
[[3], D]

By default, if you don’t specify a size, the groupTuple operator will not emit any groups until all inputs have been received. If possible, you should always try to specify the number of expected elements in each group using the size option, so that each group can be emitted as soon as it’s ready. In cases where the size of each group varies based on the grouping key, you can use the built-in groupKey() function, which allows you to define a different expected size for each group:

chr_frequency = ["chr1": 2, "chr2": 3]

Channel.of(
        ['region1', 'chr1', '/path/to/region1_chr1.vcf'],
        ['region2', 'chr1', '/path/to/region2_chr1.vcf'],
        ['region1', 'chr2', '/path/to/region1_chr2.vcf'],
        ['region2', 'chr2', '/path/to/region2_chr2.vcf'],
        ['region3', 'chr2', '/path/to/region3_chr2.vcf']
    )
    .map { region, chr, vcf -> tuple( groupKey(chr, chr_frequency[chr]), vcf ) }
    .groupTuple()
    .view()

[chr1, [/path/to/region1_chr1.vcf, /path/to/region2_chr1.vcf]]
[chr2, [/path/to/region1_chr2.vcf, /path/to/region2_chr2.vcf, /path/to/region3_chr2.vcf]]

Available options:

by

The zero-based index of the element to use as the grouping key. Can also be a list of indices, e.g. by: [0,2] (default: [0]).

remainder

When true, incomplete tuples (i.e. groups with less than size items) are emitted as partial groups, otherwise they are discarded (default: false). This option can only be used with size.

size

The required number of items for each group. When a group reaches the required size, it is emitted.

sort

Defines the sorting criteria for the grouped items. Can be one of the following values:

false: No sorting is applied (default).
true: Order the grouped items by the item’s natural ordering i.e. numerical for number, lexicographic for string, etc. See the Java documentation for more information.
'hash': Order the grouped items by the hash number associated to each entry.
'deep': Similar to the previous, but the hash number is created on actual entries content e.g. when the item is a file, the hash is created on the actual file content.
A custom sorting criteria used to order the nested list elements of each tuple. It can be a Closure or a Comparator object.

ifEmpty

Returns: value channel

The ifEmpty operator emits a source channel, or a default value if the source channel is empty (doesn’t emit any value):

Channel .of(1, 2, 3) .ifEmpty('Hello') .view()

1
2
3

Channel .empty() .ifEmpty('Hello') .view()

Hello

The default value can also be a closure, in which case the closure is evaluated and the result is emitted when the source channel is empty.

join

Returns: queue channel

The join operator emits the inner product of two source channels using a matching key.

To be more precise, the operator transforms a sequence of tuples like (K, V1, V2, ..) and (K, W1, W1, ..) into a sequence of tuples like (K, V1, V2, .., W1, W2, ..). It is equivalent to an inner join in SQL, or an outer join when remainder is true.

For example:

left  = Channel.of( ['X', 1], ['Y', 2], ['Z', 3], ['P', 7] )
right = Channel.of( ['Z', 6], ['Y', 5], ['X', 4] )

left.join(right).view()

[Z, 3, 6]
[Y, 2, 5]
[X, 1, 4]

By default, the first element of each item is used as the key. The by option can be used to specify a different index, or list of indices.

By default, unmatched items are discarded. The remainder option can be used to emit them at the end:

left  = Channel.of( ['X', 1], ['Y', 2], ['Z', 3], ['P', 7] )
right = Channel.of( ['Z', 6], ['Y', 5], ['X', 4] )

left.join(right, remainder: true).view()

[Y, 2, 5]
[Z, 3, 6]
[X, 1, 4]
[P, 7, null]

Available options:

by: The zero-based index of each item to use as the matching key. Can also be a list of indices, e.g. by: [0, 2] (default: [0]).
failOnDuplicate: When true, an error is reported when the operator receives multiple items from the same channel with the same key (default: true if strict mode is enabled, false otherwise).
failOnMismatch: When true, an error is reported when the operator receives an item from one channel for which there no matching item from the other channel (default: true if strict mode is enabled, false otherwise). This option cannot be used with remainder.
remainder: When true, unmatched items are emitted at the end, otherwise they are discarded (default: false).

last

Returns: value channel

The last operator emits the last item from a source channel:

Channel.of( 1, 2, 3, 4, 5, 6 )
    .last()
    .view()

map

Returns: queue channel

The map operator applies a mapping function to each item from a source channel:

Channel.of( 1, 2, 3, 4, 5 )
    .map { it * it }
    .view()

max

Returns: value channel

The max operator emits the item with the greatest value from a source channel:

Channel.of( 8, 6, 2, 5 )
    .max()
    .view { "Max value is $it" }

Max value is 8

An optional closure can be used to control how the items are compared. The closure can be a mapping function, which transforms each item before it is compared, or a comparator function, which defines how to compare two items more generally.

The following examples show how to find the longest string in a channel:

// mapping function
Channel.of( "hello", "hi", "hey" )
    .max { it.size() }
    .view()

hello

// comparator function
Channel.of( "hello", "hi", "hey" )
    .max { a, b -> a.size() <=> b.size() }
    .view()

hello

merge

Returns: queue channel

The merge operator joins the items from two or more channels into a new channel:

odds  = Channel.of(1, 3, 5, 7, 9)
evens = Channel.of(2, 4, 6)

odds
    .merge( evens )
    .view()

[1, 2]
[3, 4]
[5, 6]

An optional closure can be used to control how two items are merged:

odds  = Channel.of(1, 3, 5, 7, 9)
evens = Channel.of(2, 4, 6)

odds
    .merge( evens ) { a, b -> tuple(b*b, a) }
    .view()

[16, 3]
[36, 5]
[4, 1]

Danger

In general, the use of the merge operator is discouraged. Processes and channel operators are not guaranteed to emit items in the order that they were received, as they are executed concurrently. Therefore, if you try to merge output channels from different processes, the resulting channel may be different on each run, which will cause resumed runs to not work properly.

You should always use a matching key (e.g. sample ID) to merge multiple channels, so that they are combined in a deterministic way. For this purpose, you can use the join operator.

min

Returns: value channel

The min operator emits the item with the lowest value from a source channel:

Channel.of( 8, 6, 2, 5 )
    .min()
    .view { "Min value is $it" }

Min value is 2

An optional closure can be used to control how the items are compared. The closure can be a mapping function, which transforms each item before it is compared, or a comparator function, which defines how to compare two items more generally.

The following examples show how to find the shortest string in a channel:

// mapping function
Channel.of( "hello", "hi", "hey" )
    .min { it.size() }
    .view()

hi

// comparator function
Channel.of( "hello", "hi", "hey" )
    .min { a, b -> a.size() <=> b.size() }
    .view()

hi

mix

Returns: queue channel

The mix operator emits the items from two or more source channels into a single output channel:

c1 = Channel.of( 1, 2, 3 )
c2 = Channel.of( 'a', 'b' )
c3 = Channel.of( 'z' )

c1.mix(c2, c3).view()

1
2
3
a
b
z

The items in the mixed output channel may appear in any order, regardless of which source channel they came from. Thus, the previous example could also output the following:

z
1
a
2
b
3

multiMap

New in version 19.11.0-edge.

Returns: map of queue channels

The multiMap operator applies a set of mapping functions to a source channel, producing a separate output channel for each mapping function.

The multi-map criteria is a closure that defines, for each output channel, a label followed by a mapping expression.

For example:

Channel.of( 1, 2, 3, 4 )
    .multiMap { it ->
        foo: it + 1
        bar: it * it
    }
    .set { result }

result.foo.view { "foo $it" }
result.bar.view { "bar $it" }

foo 2
foo 3
foo 4
foo 5
bar 1
bar 4
bar 9
bar 16

Multiple labels can share the same mapping expression using the following shorthand:

Channel.of( 1, 2, 3 )
    .multiMap { it -> foo: bar: it }
    .set { result }

result.foo.view { "foo $it" }
result.bar.view { "bar $it" }

bar 1
bar 2
bar 3
foo 1
foo 2
foo 3

The above example creates two channels as before, but now they both receive the same items.

You can use the multiMapCriteria() method to create a multi-map criteria as a variable that can be passed as an argument to any number of multiMap operations, as shown below:

def criteria = multiMapCriteria {
    small: [it, it < 10]
    large: [it, it > 10]
}

Channel.of(1, 2, 30).multiMap(criteria).set { ch1 }
Channel.of(10, 20, 1).multiMap(criteria).set { ch2 }

ch1.small.view { v, is_small -> "ch1: $v is small: $is_small" }
ch1.large.view { v, is_large -> "ch1: $v is large: $is_large" }
ch2.small.view { v, is_small -> "ch2: $v is small: $is_small" }
ch2.large.view { v, is_large -> "ch2: $v is large: $is_large" }

Note

If you use multiMap to split a tuple or map into multiple channels, it is recommended that you retain a matching key (e.g. sample ID) with each new channel, so that you can re-combine these channels later on if needed. In general, you should not expect to be able to merge channels correctly without a matching key, due to the concurrent nature of Nextflow pipelines.

randomSample

Returns: queue channel

The randomSample operator emits a randomly-selected subset of items from a source channel:

Channel.of( 1..100 )
    .randomSample( 10 )
    .view()

The above snippet will print 10 randomly-selected numbers between 1 and 100 (without replacement).

An optional second parameter can be used to set the initial seed for the random number generator, which ensures that the randomSample operator produces the same pseudo-random sequence across runs:

Channel.of( 1..100 )
    .randomSample( 10, 234 )
    .view()

The above example will print 10 randomly-selected numbers between 1 and 100 (without replacement). Each subsequent script execution will produce the same sequence.

reduce

Returns: value channel

The reduce operator applies an accumulator function sequentially to each item in a source channel, and emits the final accumulated value. The accumulator function takes two parameters – the accumulated value and the i-th emitted item – and it should return the accumulated result, which is passed to the next invocation with the i+1-th item. This process is repeated for each item in the source channel.

For example:

Channel.of( 1, 2, 3, 4, 5 )
    .reduce { a, b ->
        println "a: $a b: $b"
        a + b
    }
    .view { "result = $it" }

a: 1 b: 2
a: 3 b: 3
a: 6 b: 4
a: 10 b: 5
result = 15

By default, the first item is used as the initial accumulated value. You can optionally specify a different initial value as shown below:

Channel.of( 1, 2, 3, 4, 5 )
    .reduce( 'result:' ) { accum, v ->
        println accum
        accum + ' ' + v
    }
    .view { "final $it" }

result:
result: 1
result: 1 2
result: 1 2 3
result: 1 2 3 4
final result: 1 2 3 4 5

set

Returns: nothing

The set operator assigns a source channel to a variable, whose name is specified as a closure parameter:

Channel.of(10, 20, 30).set { my_channel }

Using set is semantically equivalent to assigning a variable:

my_channel = Channel.of(10, 20, 30)

splitCsv

Returns: queue channel

The splitCsv operator parses and splits CSV-formatted text from a source channel into records, or groups of records with a given size.

For example:

Channel.of( '10,20,30\n70,80,90' )
    .splitCsv()
    .view { row -> "${row[0]} - ${row[1]} - ${row[2]}" }

10 - 20 - 30
70 - 80 - 90

The above example shows hows CSV text is parsed and split into individual rows, where each row is simply a list of columns.

When the CSV begins with a header line defining the column names, and the header option is true, each row is returned as a map instead:

Channel.of( 'alpha,beta,gamma\n10,20,30\n70,80,90' )
    .splitCsv( header: true )
    .view { row -> "${row.alpha} - ${row.beta} - ${row.gamma}" }

10 - 20 - 30
70 - 80 - 90

The header option can also just be a list of columns:

Channel.of( 'alpha,beta,gamma\n10,20,30\n70,80,90' )
    .splitCsv( header: ['col1', 'col2', 'col3'], skip: 1 )
    .view { row -> "${row.col1} - ${row.col2} - ${row.col3}" }

10 - 20 - 30
70 - 80 - 90

Available options:

by: When specified, group rows into chunks with the given size (default: none).
charset: Parse the content with the specified charset, e.g. UTF-8. See the list of standard charsets for available options.
decompress: When true, decompress the content using the GZIP format before processing it (default: false). Files with the .gz extension are decompressed automatically.
elem: The index of the element to split when the source items are lists or tuples (default: first file object or first element).
header: When true, the first line is used as the columns names (default: false). Can also be a list of columns names.
limit: Limits the number of records to retrieve for each source item (default: no limit).
quote: The character used to quote values (default: '' or "").
sep: The character used to separate values (default: ,)
skip: Number of lines to ignore from the beginning when parsing the CSV text (default: 0).
strip: When true, remove leading and trailing blanks from values (default: false).

splitFasta

Returns: queue channel

The splitFasta operator splits FASTA-formatted text from a source channel into individual sequences.

The by option can be used to group sequences into chunks of a given size. The following example shows how to read a FASTA file and split it into chunks of 10 sequences each:

Channel
     .fromPath('misc/sample.fa')
     .splitFasta( by: 10 )
     .view()

Warning

Chunks are stored in memory by default. When splitting large files, specify file: true to save the chunks into files in order to avoid running out of memory. See the list of options below for details.

The record option can be used to split FASTA content into records instead of text chunks. Each record is a map that allows you to access the FASTA sequence data with ease. For example:

Channel
     .fromPath('misc/sample.fa')
     .splitFasta( record: [id: true, seqString: true] )
     .filter { record -> record.id =~ /^ENST0.*/ }
     .view { record -> record.seqString }

The above example loads the misc/sample.fa file, splits it into records containing the id and seqString fields (i.e. the sequence id and the sequence data), filters records by their ID, and finally prints the sequence string of each record.

Available options:

by

Defines the number of sequences in each chunk (default: 1).

charset

Parse the content with the specified charset, e.g. UTF-8. See the list of standard charsets for available options.

compress

When true, resulting file chunks are GZIP compressed (default: false). The .gz suffix is automatically added to chunk file names.

decompress

When true, decompress the content using the GZIP format before processing it (default: false). Files with the .gz extension are decompressed automatically.

elem

The index of the element to split when the source items are lists or tuples (default: first file object or first element).

file

When true, saves each split to a file. Use a string instead of true value to create split files with a specific name (split index number is automatically added). Finally, set this attribute to an existing directory, in order to save the split files into the specified directory.

limit

Limits the number of sequences to retrieve for each source item (default: no limit).

record

Parse each entry in the FASTA file into a record. The following fields are available:

id: The FASTA sequence identifier, i.e. the word following the > symbol up to the first blank or newline character
header: The first line in a FASTA sequence without the > character
desc: The text in the FASTA header following the ID value
text: The complete FASTA sequence including the header
seqString: The sequence data as a single-line string, i.e. containing no newline characters
sequence: The sequence data as a multi-line string, i.e. always ending with a newline character
width: Define the length of a single line when the sequence field is used, after which the sequence data continues on a new line.

size

Defines the size of the expected chunks as a memory unit, e.g. 1.MB.

splitFastq

Returns: queue channel

The splitFasta operator splits FASTQ formatted text from a source channel into individual sequences.

The by option can be used to group sequences into chunks of a given size. The following example shows how to read a FASTQ file and split it into chunks of 10 sequences each:

Channel
    .fromPath('misc/sample.fastq')
    .splitFastq( by: 10 )
    .view()

Warning

Chunks are stored in memory by default. When splitting large files, specify file: true to save the chunks into files in order to avoid running out of memory. See the list of options below for details.

The record option can be used to split FASTQ content into records instead of text chunks. Each record is a map that allows you to access the FASTQ sequence data with ease. For example:

Channel
    .fromPath('misc/sample.fastq')
    .splitFastq( record: true )
    .view { record -> record.readHeader }

The pe option can be used to split paired-end FASTQ files. The source channel must emit tuples containing the file pairs. For example:

Channel
    .fromFilePairs('/my/data/SRR*_{1,2}.fastq', flat: true)
    .splitFastq(by: 100_000, pe: true, file: true)
    .view()

Note

Channel.fromFilePairs() requires the flat: true option in order to emit the file pairs as separate elements in the produced tuples.

Note

This operator assumes that the order of the paired-end reads correspond with each other and that both files contain the same number of reads.

Available options:

by

Defines the number of sequences in each chunk (default: 1).

charset

Parse the content with the specified charset, e.g. UTF-8. See the list of standard charsets for available options.

compress

When true, resulting file chunks are GZIP compressed (default: false). The .gz suffix is automatically added to chunk file names.

decompress

When true, decompress the content using the GZIP format before processing it (default: false). Files with the .gz extension are decompressed automatically.

elem

The index of the element to split when the source items are lists or tuples (default: first file object or first element).

file

When true, saves each split to a file. Use a string instead of true value to create split files with a specific name (split index number is automatically added). Finally, set this attribute to an existing directory, in order to save the split files into the specified directory.

limit

Limits the number of sequences to retrieve for each source item (default: no limit).

pe

When true, splits paired-end read files. Items emitted by the source channel must be tuples with the file pairs.

record

Parse each entry in the FASTQ file into a record. The following fields are available:

readHeader: Sequence header (without the @ prefix)
readString: The raw sequence data
qualityHeader: Base quality header (it may be empty)
qualityString: Quality values for the sequence

splitJson

Returns: queue channel

The splitJson operator splits JSON formatted text from a source channel into individual records.

If the source item is a JSON array, each element of the array will be emitted:

// Example with a JSON array
Channel.of('[1, null, ["A", {}], true]')
    .splitJson()
    .view{"Item: ${it}"}

Item: 1
Item: null
Item: [A, [:]]
Item: true

If the source item is a JSON object, each key-value pair will be emitted as a map with the properties key and value:

// Example with a JSON object
Channel.of('{"A": 1, "B": [1, 2, 3], "C": {"D": null}}')
    .splitJson()
    .view{"Item: ${it}"}

Item: [key:A, value:1]
Item: [key:B, value:[1, 2, 3]]
Item: [key:C, value:[D:null]]

The path option can be used to query a section of the JSON document to parse and split:

Channel.of('{"A": 1, "B": [2, 3, {"C": {"D": null, "E": 4, "F": 5}}]}')
    .splitJson(path: 'B[2].C')
    .view{"Item: ${it}"}

Item: [key:D, value:null]
Item: [key:E, value:4]
Item: [key:F, value:5]

Available options:

limit: Limits the number of records to retrieve for each source item (default: no limit).
path: Defines a query for a section of each source item to parse and split. The expression should be a path similar to JSONPath. The empty string is the document root (default). An integer in brackets is a zero-based index in a JSON array. A string preceded by a dot . is a key in a JSON object.

splitText

Returns: queue channel

The splitText operator splits multi-line text content from a source channel into chunks of N lines:

Channel
    .fromPath('/some/path/*.txt')
    .splitText()
    .view()

The above example loads a collection of text files, splits the content of each file into individual lines, and prints each line.

The by option can be used to emit chunks of N lines:

Channel
    .fromPath('/some/path/*.txt')
    .splitText( by: 10 )
    .subscribe {
        print it;
        print "--- end of the chunk ---\n"
    }

An optional closure can be used to transform each text chunk produced by the operator. The following example shows how to split text files into chunks of 10 lines and transform them to uppercase letters:

Channel
    .fromPath('/some/path/*.txt')
    .splitText( by: 10 ) { it.toUpperCase() }
    .view()

Note

Text chunks returned by the splitText operator are always terminated by a \n newline character.

Available options:

by: Defines the number of lines in each chunk (default: 1).
charset: Parse the content with the specified charset, e.g. UTF-8. See the list of standard charsets for available options.
compress: When true, resulting file chunks are GZIP compressed (default: false). The .gz suffix is automatically added to chunk file names.
decompress: When true, decompresses the content using the GZIP format before processing it (default: false). Files with the .gz extension are decompressed automatically.
elem: The index of the element to split when the source items are lists or tuples (default: first file object or first element).
file: When true, saves each split to a file. Use a string instead of true value to create split files with a specific name (split index number is automatically added). Finally, set this attribute to an existing directory, in order to save the split files into the specified directory.
keepHeader: Parses the first line as header and prepends it to each emitted chunk (default: false).
limit: Limits the number of lines to retrieve for each source item (default: no limit).

sum

Returns: value channel

The sum operator emits the sum of all items in a source channel:

Channel.of( 8, 6, 2, 5 )
    .sum()
    .view { "The sum is $it" }

The sum is 21

An optional closure can be used to transform each item before it is added to the sum:

Channel.of( 4, 1, 7, 5 )
    .sum { it * it }
    .view { "Square: $it" }

Square: 91

take

Returns: queue channel

The take operator takes the first N items from a source channel:

Channel.of( 1..10 )
    .take( 3 )
    .view()

1
2
3

Tip

Specifying a size of -1 causes the operator to take all values.

tap

Returns: queue channel

The tap operator assigns a source channel to a variable, and emits the source channel. It is a useful way to extract intermediate output channels from a chain of operators. For example:

Channel.of( 'a', 'b', 'c' )
    .tap { log1 }
    .map { it * 2 }
    .tap { log2 }
    .map { it.toUpperCase() }
    .view { "Result: $it" }

log1.view { "Log 1: $it" }
log2.view { "Log 2: $it" }

Result: AA
Result: BB
Result: CC
Log 1: a
Log 1: b
Log 1: c
Log 2: aa
Log 2: bb
Log 2: cc

toInteger

Returns: queue channel

The toInteger operator converts string values from a source channel to integer values:

Channel.of( '1', '7', '12' )
    .toInteger()
    .sum()
    .view()

Note

toInteger is equivalent to:

map { it -> it as Integer }

Note

You can also use toLong, toFloat, and toDouble to convert to other numerical types.

toList

Returns: value channel

The toList operator collects all the items from a source channel into a list and emits the list as a single item:

Channel.of( 1, 2, 3, 4 )
    .toList()
    .view()

[1, 2, 3, 4]

Note

There are two main differences between toList and collect:

When there is no input, toList emits an empty list whereas collect emits nothing.
By default, collect flattens list items by one level.

In other words, toList is equivalent to:

collect(flat: false).ifEmpty([])

toSortedList

Returns: value channel

The toSortedList operator collects all the items from a source channel into a sorted list and emits the list as a single item:

Channel.of( 3, 2, 1, 4 )
    .toSortedList()
    .view()

[1, 2, 3, 4]

An optional closure can be used to control how items are compared when sorting. For example, to sort tuples by their second element in descending order:

Channel.of( ['homer', 5], ['bart', 2], ['lisa', 10], ['marge', 3], ['maggie', 7] )
    .toSortedList { a, b -> b[1] <=> a[1] }
    .view()

[[lisa, 10], [maggie, 7], [homer, 5], [marge, 3], [bart, 2]]

Note

toSortedList is equivalent to:

collect(flat: false, sort: true).ifEmpty([])

transpose

Returns: queue channel

The transpose operator “transposes” each tuple from a source channel by flattening any nested list in each tuple, emitting each nested item separately.

To be more precise, the operator transforms a sequence of tuples like (K, list(V), list(W), ..) into a sequence of tuples like (K, V, W, ..).

For example:

Channel.of(
        [1, ['A', 'B', 'C']],
        [2, ['C', 'A']],
        [3, ['B', 'D']]
    )
    .transpose()
    .view()

[1, A]
[1, B]
[1, C]
[2, C]
[2, A]
[3, B]
[3, D]

If each source item has more than two elements, these will be flattened by the first element in the item, and a new item will be emitted only when it is complete:

Channel.of(
        [1, [1], ['A']],
        [2, [1, 2], ['B', 'C']],
        [3, [1, 2, 3], ['D', 'E']]
    )
    .transpose()
    .view()

[1, 1, A]
[2, 1, B]
[2, 2, C]
[3, 1, D]
[3, 2, E]

The remainder option can be used to emit any incomplete items:

Channel.of(
        [1, [1], ['A']],
        [2, [1, 2], ['B', 'C']],
        [3, [1, 2, 3], ['D', 'E']]
    )
    .transpose(remainder: true)
    .view()

[1, 1, A]
[2, 1, B]
[2, 2, C]
[3, 1, D]
[3, 2, E]
[3, 3, null]

Available options:

by: The zero-based index of the element to be transposed. Can also be a list of indices, e.g. by: [0,2]. By default, every list element is transposed.
remainder: When true, incomplete tuples are emitted with null values for missing elements, otherwise they are discarded (default: false).

unique

Returns: queue channel

The unique operator emits the unique items from a source channel:

Channel.of( 1, 1, 2, 2, 2, 3, 1, 1, 2, 2, 3 )
    .unique()
    .view()

1
2
3

An optional closure can be used to transform each item before it is evaluated for uniqueness:

Channel.of( 1, 1, 2, 2, 2, 3, 1, 1, 2, 4, 6 )
    .unique { it % 2 }
    .view()

1
2

Note

The difference between unique and distinct is that unique removes all duplicate values, whereas distinct removes only consecutive duplicate values. As a result, unique must process the entire source channel before it can emit anything, whereas distinct can emit each value immediately.

until

Returns: queue channel

The until operator emits each item from a source channel until a stopping condition is satisfied:

Channel.of( 3, 2, 1, 5, 1, 5 )
    .until { it == 5 }
    .view()

3
2
1

view

Returns: queue channel

The view operator prints each item from a source channel to standard output:

Channel.of(1, 2, 3).view()

1
2
3

An optional closure can be used to transform each item before it is printed:

Channel.of(1, 2, 3)
    .map { it -> [it, it*it] }
    .view { num, sqr -> "The square of $num is $sqr" }

The square of 1 is 1
The square of 2 is 4
The square of 3 is 9

The view operator also emits every item that it receives, allowing it to be chained with other operators.

Available options:

newLine: Print each item to a separate line (default: true).

Operators

branch

buffer

collate

collect

collectFile

combine

concat

count

countFasta

countFastq

countJson

countLines

cross

distinct

dump

filter

first

flatMap

flatten

groupTuple

ifEmpty

join

last

map

max

merge

min

mix

multiMap

randomSample

reduce

set

splitCsv

splitFasta

splitFastq

splitJson

splitText

subscribe

sum

take

tap

toInteger

toList

toSortedList

transpose

unique

until

view