It’s surprising how often this is implemented incorrectly.
What we’re looking for is a substring in the mathematical sense.
In mathematics, a sequence is an enumerated collection of objects in which repetitions are allowed and order matters.
In mathematics, a subsequence of a given sequence is a sequence that can be derived from the given sequence by deleting some or no elements without changing the order of the remaining elements.
A subsequence which consists of a consecutive run of elements from the original sequence, such as ⟨ B, C, D ⟩ from ⟨ A, B, C, D, E, F ⟩ is a substring.
Note that a “string”, here, can consist of any element and is not limited to Unicode code-point sequences.
Effectively all previous answers have one of many possible flaws:
- The string concatenation approach (
array1.toString().includes(array2.toString())
) fails when your array elements have commas. (Example: [ "a", "b" ]
does not contain [ "a,b" ]
).
- Some implementations check beyond array bounds. (Example:
[ "3" ]
does not contain [ "3", undefined ]
, just because array[1]
reports undefined
for both).
- Some implementations fail to handle repetition correctly.
- Some implementations aren’t checking for substrings (in the mathematical sense) correctly, but for subsets or subsequences or something else.
- Some implementations don’t account for the empty array. The empty string is the substring of every string.
Check if an array constitutes a “substring” of another array
Right off the bat, this handles the empty array correctly.
Then, it builds a list of candidate starting indexes by matching against the first element of the potential subarray.
Find the first candidate where every element of the slice matches index by index with the full array, offset by the candidate starting index.
The checked index also has to exist within the full array, hence Object.hasOwn
.
const isSubArray = (full, slice) => {
if(slice.length === 0){
return true;
}
const candidateIndexes = full
.map((element, fullIndex) => ({
matched: element === slice[0],
fullIndex
}))
.filter(({ matched }) => matched),
found = candidateIndexes
.find(({ fullIndex }) => slice.every((element, sliceIndex) => Object.hasOwn(full, fullIndex + sliceIndex) && element === full[fullIndex + sliceIndex]));
return Boolean(found);
};
console.log(isSubArray([], []) === true);
console.log(isSubArray([ 0 ], []) === true);
console.log(isSubArray([ 0, 1, 2 ], [ 1, 2 ]) === true);
console.log(isSubArray([ 0, 1, 1, 2 ], [ 0, 1, 2 ]) === false);
console.log(isSubArray([ 2, 1 ], [ 1, 2 ]) === false);
console.log(isSubArray([ 1, 2, 3 ], [ 2, 3, undefined ]) === false);
console.log(isSubArray([ 0, 1, 1, 2, 3 ], [ 1, 1, 2 ]) === true);
console.log(isSubArray([ 0, 1, 1, 2, 3 ], [ 1, 2 ]) === true);
console.log(isSubArray([ 0, 1, 1, 2, 3 ], [ 0, 1, 1, 1 ]) === false);
console.log(isSubArray([ "a", "b" ], [ "a,b" ]) === false);
.as-console-wrapper { max-height: 100% !important; top: 0; }
This has quadratic complexity, yes.
There might be more efficient implementations using Trees or Ropes.
You might also want to research some efficient substring search algorithms and try to apply them to this problem.
Get the index of the found “substring”, or -1
if not found
It’s basically the same code, but with return true;
replaced by return 0;
, and return Boolean(found);
replaced by return found?.fullIndex ?? -1;
.
const findSubArrayIndex = (full, slice) => {
if(slice.length === 0){
return 0;
}
const candidateIndexes = full
.map((element, fullIndex) => ({
matched: element === slice[0],
fullIndex
}))
.filter(({ matched }) => matched),
found = candidateIndexes
.find(({ fullIndex }) => slice.every((element, sliceIndex) => Object.hasOwn(full, fullIndex + sliceIndex) && element === full[fullIndex + sliceIndex]));
return found?.fullIndex ?? -1;
};
console.log(findSubArrayIndex([], []) === 0);
console.log(findSubArrayIndex([ 0 ], []) === 0);
console.log(findSubArrayIndex([ 0, 1, 2 ], [ 1, 2 ]) === 1);
console.log(findSubArrayIndex([ 0, 1, 1, 2 ], [ 0, 1, 2 ]) === -1);
console.log(findSubArrayIndex([ 2, 1 ], [ 1, 2 ]) === -1);
console.log(findSubArrayIndex([ 1, 2, 3 ], [ 2, 3, undefined ]) === -1);
console.log(findSubArrayIndex([ 0, 1, 1, 2, 3 ], [ 1, 1, 2 ]) === 1);
console.log(findSubArrayIndex([ 0, 1, 1, 2, 3 ], [ 1, 2 ]) === 2);
console.log(findSubArrayIndex([ 0, 1, 1, 2, 3 ], [ 0, 1, 1, 1 ]) === -1);
console.log(findSubArrayIndex([ "a", "b" ], [ "a,b" ]) === -1);
.as-console-wrapper { max-height: 100% !important; top: 0; }
Semi-acceptable alternative: JSON
JSON-encoding both arrays might be a viable strategy as well.
Here, the surrounding [
…]
of the potential subarray need to be removed, then an includes
will tell you if the JSON string is included in the other JSON string.
This works — as opposed to the simple string concatenation or join
approach — because JSON has delimiters that cannot appear verbatim in the encoded elements; if they do appear in the original elements, they’d be correctly escaped.
The caveat is that this won’t work for values that are not encodable in JSON.
const isSubArray = (full, slice) => JSON.stringify(full)
.includes(JSON.stringify(slice).replaceAll(/^\[|\]$/g, ""));
console.log(isSubArray([], []) === true);
console.log(isSubArray([ 0 ], []) === true);
console.log(isSubArray([ 0, 1, 2 ], [ 1, 2 ]) === true);
console.log(isSubArray([ 0, 1, 1, 2 ], [ 0, 1, 2 ]) === false);
console.log(isSubArray([ 2, 1 ], [ 1, 2 ]) === false);
console.log(isSubArray([ 1, 2, 3 ], [ 2, 3, undefined ]) === false);
console.log(isSubArray([ 0, 1, 1, 2, 3 ], [ 1, 1, 2 ]) === true);
console.log(isSubArray([ 0, 1, 1, 2, 3 ], [ 1, 2 ]) === true);
console.log(isSubArray([ 0, 1, 1, 2, 3 ], [ 0, 1, 1, 1 ]) === false);
console.log(isSubArray([ "a", "b" ], [ "a,b" ]) === false);
.as-console-wrapper { max-height: 100% !important; top: 0; }