Database key compression

1. A method, comprising:

receiving a request for a database record stored on a storage medium, wherein the database record includes a compressed database key having information that specifies:

a location of a reference database key; and

an amount of similarity between the reference database key and an uncompressed version of the compressed database key;

accessing the reference database key based on the location specified by the compressed database key;

decompressing the compressed database key based on the reference database key and the amount of similarity; and

returning the database record with the decompressed database key.

2. The method of claim 1 , wherein the database record is written to a data block that stores a set of database records, and wherein the reference database key is stored in one of a plurality of index blocks that comprise an index associated with the data block.

3. The method of claim 2 , wherein the information includes a first portion that specifies the index block and a second portion that specifies an entry within the index block that stores the reference database key.

4. The method of claim 1 , wherein the amount of similarity indicates an amount of bytes that are contiguously shared, starting from an initial byte position, between the reference database key and the uncompressed version of the compressed database key.

5. The method of claim 1 , further comprising:

compressing the uncompressed version of the compressed database key, including:

based on determining that the database record is the logical first database record in a set of database records, adding the reference database key to a set of reference database keys that is maintained by a computer system that performs the method.

6. The method of claim 1 , wherein an uncompressed version of another database key shares a different amount of similarity with the reference database key than the uncompressed version of the compressed database key.

7. The method of claim 1 , wherein the database record is included in a set of database records having database keys that define a particular database key-sorted order, and wherein compressed versions of the database keys defined the same particular database key-sorted order.

8. A non-transitory computer readable medium having program instructions stored thereon that are capable of causing a computer system to perform operations comprising:

receiving a request for a database record stored on a storage medium, wherein the database record includes a compressed database key having information that specifies:

a location of a reference database key; and

an amount of similarity between the reference database key and an uncompressed version of the compressed database key;

accessing the reference database key based on the location specified by the compressed database key;

decompressing the compressed database key based on the reference database key and the amount of similarity; and

returning the database record with the decompressed database key.

9. The medium of claim 8 , wherein the operations further comprise:

writing the database record to a data block that stores a set of database records; and

writing the reference database key to one of a plurality of index blocks that define an index associated with the data block.

10. The medium of claim 9 , wherein the information includes a first portion that specifies the index block and a second portion that specifies an entry within the index block that stores the reference database key.

11. The medium of claim 8 , wherein the amount of similarity indicates a number of characters that are contiguously shared, from an initial character position, between the reference database key and the uncompressed version of the compressed database key.

12. The medium of claim 8 , wherein the operations further comprise:

receiving a request to store the database record on the storage medium;

compressing the uncompressed version of the database key by replacing a portion of the database key with data that identifies the location of the reference database key and the amount of similarity; and

storing the database record with the compressed database key.

13. The medium of claim 8 , wherein the database record is written to a file of a particular level of a log-structured merge-tree (LSM tree), wherein the LSM tree includes multiple levels storing files, and wherein the reference database key is specific to the particular level.

14. The medium of claim 8 , wherein the operations further comprise:

receiving a request for a different database record;

accessing the reference database key;

decompressing a database key of the different record based on the reference database key and a different amount of similarity; and

returning the different database record.

15. A system, comprising:

at least one processor; and

memory having program instructions stored thereon that are executable by the at least one processor to cause the system to perform operations comprising:

receiving a request for a database record stored on a storage medium, wherein the database record includes a compressed database key having information that specifies:

a location of a reference database key; and

an amount of similarity between the reference database key and an uncompressed version of the compressed database key;

accessing the reference database key based on the location specified by the compressed database key;

decompressing the compressed database key based on the reference database key and the amount of similarity; and

returning the database record with the decompressed database key.

16. The system of claim 15 , wherein the operations further comprise:

writing the database record to a data block and the reference database key to an index block that is separate from the data block.

17. The system of claim 16 , wherein the operations further comprise:

prior to receiving the request, compressing the uncompressed version of the database key based on the reference database key, wherein the compressing includes replacing a portion of the database key with a first portion that specifies the index block and a second portion that specifies an entry within the index block that stores the reference database key.

18. The system of claim 16 , wherein the writing includes storing, with the reference database key, pointer information that identifies the data block that stores the database record.

19. The system of claim 15 , wherein the amount of similarity indicates an amount of bytes that are contiguously shared, starting from an initial byte position, between the reference database key and the uncompressed version of the compressed database key.

20. The system of claim 15 , wherein the operations further comprise:

receiving a request for a different database record;

accessing the reference database key;

decompressing a database key of the different record based on the reference database key and the amount of similarity; and

returning the different database record.