I don’t think there is any limit set by Elaticsearch or Lucene explicitly. The limit you might hit, though, is the one set in place by the JDK.
To prove my statement above, I looked at the source code of Elasticsearch:
-
when the request comes in there is a parser that parses the array of ids. All it’s using is an
ArrayList
. This is then passed along to Lucene, which in turn it’s using a List. -
this is the Lucene TermsFilter class (line #84) that gets the list of IDS from Elasticsearch within a List.
-
source code of
ArrayList
class from Oracle JDK 1.7.0_67:
/**
* The maximum size of array to allocate.
* Some VMs reserve some header words in an array.
* Attempts to allocate larger arrays may result in
* OutOfMemoryError: Requested array size exceeds VM limit
*/
private static final int MAX_ARRAY_SIZE = Integer.MAX_VALUE - 8;
/**
* Increases the capacity to ensure that it can hold at least the
* number of elements specified by the minimum capacity argument.
*
* @param minCapacity the desired minimum capacity
*/
private void grow(int minCapacity) {
...
if (newCapacity - MAX_ARRAY_SIZE > 0)
newCapacity = hugeCapacity(minCapacity);
...
}
private static int hugeCapacity(int minCapacity) {
if (minCapacity < 0) // overflow
throw new OutOfMemoryError();
return (minCapacity > MAX_ARRAY_SIZE) ?
Integer.MAX_VALUE :
MAX_ARRAY_SIZE;
}
And that number (Integer.MAX_VALUE - 8
) is 2147483639
. So, this would be the theoretical max size of that array.
I’ve tested locally in my ES instance an array of 150000 elements. And here comes the performance implications: of course, you would get a degrading performance the larger the array gets. In my simple test with 150k ids I got a 800 ms execution time. But, all depends on CPU, memory, load, datasize, data mapping etc etc. The best would be for you to actually test this.
UPDATED Dec. 2016: this answer applies for the Elasticsearch version in existence at the end of 2014, ie in the 1.x branch. The latest available at that time was 1.4.x.