Need memory efficient way to store tons of strings (was: HAT-Trie implementation in java)

hashable picture hashable · Feb 8, 2010 · Viewed 11.9k times · Source

I am working with a large set (5-20 million) of String keys (average length 10 chars) which I need to store in an in memory data structure that supports the following operation in constant time or near constant time:

// Returns true if the input is present in the container, false otherwise
public boolean contains(String input)

Java's Hashmap is proving to be more than satisfactory as far as throughput is concerned but is taking up a lot of memory. I am looking for a solution that is memory efficient and still supports a throughput that is decent (comparable with or nearly as good as hashing).

I don't care about the insertion/deletion times. In my application, I will be performing only insertions (only at startup time) and will subsequently only be querying the data structure using the contains method for the life of the application.

I read that the HAT-Trie data structure is closest for my needs. I am wondering if there is a library that has an implementation.

Other suggestions with pointers to implementations welcome.

Thank You.

Answer

Pascal Cuoq picture Pascal Cuoq · Feb 8, 2010

The trie seems like a very good idea for your constraints.

A "thinking outside the box" alternative:

If you can afford some probability of answering "present" for a string that is absent

EDIT: if you can afford false positives, use a Bloom filter as suggested by WizardOfOdds in the comments.

For k=1, a Bloom filter is like a hash table without the keys: each "bucket" is simply a boolean that tells if at least one input with the same hash was present. If 1% false positives is acceptable, your hash table can be as small as about 100 * 20 million bits or roughly 200 MiB. For 1 in 1000 false positives, 2GiB.

Using several hash functions instead of one can improve the false positive rate for the same amount of bits.