Entries in split (2)

How do I split a list?

If you want to split the list into lists of equal size, see this related faq on the # (take) operator.

If you want to split a string on a delimiter, use the vs (vector from scalar) function:

q)" " vs "The quick brown fox"
"The"
"quick"
"brown"
"fox"
q)", " vs "Hello, world!"
"Hello"
"world!"
q)

When its left argument is the null symbol, `, the vs function breaks apart a symbol on dots:

q)` vs `foo.bar.baz
`foo`bar`baz
q)

Those are the most common cases. We can also split a list into lists of varying length by passing a list of indexes as the left argument to the _ (cut) operator:

q)0 1 4 9 _ til 10
,0
1 2 3
4 5 6 7 8
,9
q)

To split a list of some type other than char using a delimiter is a little more complicated. We start by finding the indexes in the list that match the delimiter:

q)list: 1 2 0 3 4 0 5 6
q)delimiter: 0
q)indexes: where delimiter = list
q)indexes
2 5
q)

Now we can break list into pieces using the _ (cut) operator as above:

q)indexes _ list
0 3 4
0 5 6
q)

This is almost what we want. We'll use _/: (drop-each-right) to get rid of the delimiters:

q)1 _/: indexes _ list
3 4
5 6
q)

We can grab the first element of the result with # (take):

q)first[indexes] # list
1 2
q)

Then we can just join (using ,) the two together:

q)(enlist first[indexes] # list), 1 _/: indexes _ list
1 2
3 4
5 6
q)

Note that we must call enlist on the front of the list or else we'll get something a little different from what we intended:

q)(first[indexes] # list), 1 _/: indexes _ list
1
2
3 4
5 6
q)

Lastly, we can generalize to non-atomic types by replacing = with ~/: (match-each-right):

split: {[list; delimiter]
    indexes: where delimiter ~/: list;
    front: first[indexes] # list;
    rest: 1 _/: indexes _ list;
    : (enlist front), rest;
    }

How can I slice a vector into a series of vectors of length n?

By passing the # (take) operator a pair as its first (i.e., left) argument:

q)2 5 # til 10
0 1 2 3 4
5 6 7 8 9

When # is used in this way, it is referred to as "reshape." We like to call this, "taking a box." Just like when you apply # with an integer left argument, taking a box stops when you have taken as many items as requested, leaving the rest behind:

q)3 3 # til 10
0 1 2
3 4 5
6 7 8

Similarly, when you over-take a box, you start over from the beginning of the source:

q)4 3 # til 10
0 1 2
3 4 5
6 7 8
9 0 1

If you want to take all the items from the source, pass 0N as the first element of the box:

q)0N 2 # til 10
0 1
2 3
4 5
6 7
8 9

If you take a whole box, but the elements from the source don't fit, you end up with a short list in the last row:

q)0N 3 # til 10
0 1 2
3 4 5
6 7 8
,9
q)

See also: cut and faq on _ (drop)