Fingerprints for compressed columnar data search

1. A computer-implemented method, comprising:

compressing columnar data to be stored in an in-memory database, including:

dividing the columnar data into a plurality of data blocks; and

in response to dividing the columnar data into the plurality of data blocks, and for each particular data block in the plurality of data blocks, generating a particular fingerprint associated with the particular data block, wherein the particular fingerprint indicates a data range and one or more data gaps of the particular data block, and wherein the particular fingerprint is generated using one of a plurality of predefined fingerprint schemes based on data distribution in the columnar data;

storing the compressed columnar data and the generated fingerprints in the in-memory database;

receiving a query for the columnar data;

for each particular in-memory data block in the plurality of data blocks stored in the in-memory database:

determining whether the particular in-memory data block satisfies the query based on the particular fingerprint associated with the particular in-memory data block and on the data range and the one or more data gaps of the particular in-memory data block indicated by the particular fingerprint; and

in response to a determination that the particular in-memory data block does not satisfy the query, pruning the particular in-memory data block from the plurality of data blocks to generate an unpruned set of data blocks;

decompressing the unpruned set of data blocks; and

performing a query search on the decompressed unpruned set of data blocks for the received query.

2. The computer-implemented method of claim 1 , wherein the plurality of predefined fingerprint schemes includes at least two of a Basic Fingerprint (BF) scheme, a Unilateral Fingerprint (UF) scheme, a Selective Fingerprint (SF) scheme, a Two-Level Fingerprint (TLF) scheme, an Impartial Fingerprint (IF) scheme, and a Heterogeneous Fingerprint (HF) scheme.

3. The computer-implemented method of claim 1 , wherein the query includes a query predicate, and wherein determining whether the particular in-memory data block satisfies the query includes determining whether the particular in-memory data block satisfies the query predicate.

4. The computer-implemented method of claim 3 , wherein the query predicate is a range predicate, and wherein determining whether the particular in-memory data block satisfies the query predicate includes comparing a fingerprint associated with the particular in-memory data block and an encoding of the range predicate using a bitwise AND operation.

5. The computer-implemented method of claim 1 , wherein the determining and pruning are performed on the compressed columnar data and the generated fingerprints stored in the in-memory database.

6. The computer-implemented method of claim 1 , wherein any data block in the unpruned set of data blocks satisfies the query, and wherein decompressing the unpruned set of data blocks includes vectorizing values in the unpruned set of data blocks.

7. A non-transitory computer storage medium encoded with a computer program, the program comprising instructions that, when executed by one or more computers, cause the one or more computers to perform operations comprising:

compressing columnar data to be stored in an in-memory database, including:

dividing the columnar data into a plurality of data blocks; and

in response to dividing the columnar data into the plurality of data blocks, for each particular data block in the plurality of data blocks, generating a particular fingerprint associated with the particular data block, wherein the particular fingerprint indicates a data range and one or more data gaps of the particular data block, and wherein the particular fingerprint is generated using one of a plurality of predefined fingerprint schemes based on data distribution in the columnar data;

storing the compressed columnar data and the generated fingerprints in the in-memory database;

receiving a query for the columnar data;

for each particular in-memory data block in the plurality of data blocks stored in the in-memory database:

determining whether the particular in-memory data block satisfies the query based on the particular fingerprint associated with the particular in-memory data block and on the data range and the one or more data gaps of the particular in-memory data block indicated by the particular fingerprint; and

in response to a determination that the particular in-memory data block does not satisfy the query, pruning the particular in-memory data block from the plurality of data blocks to generate an unpruned set of data blocks;

decompressing the unpruned set of data blocks; and

performing a query search on the decompressed unpruned set of data blocks for the received query.

8. The non-transitory computer storage medium of claim 7 , wherein the plurality of predefined fingerprint schemes includes at least two of a Basic Fingerprint (BF) scheme, a Unilateral Fingerprint (UF) scheme, a Selective Fingerprint (SF) scheme, a Two-Level Fingerprint (TLF) scheme, an Impartial Fingerprint (IF) scheme, and a Heterogeneous Fingerprint (HF) scheme.

9. The non-transitory computer storage medium of claim 7 , wherein the query includes a query predicate, and wherein determining whether the particular in-memory data block satisfies the query includes determining whether the particular in-memory data block satisfies the query predicate.

10. The non-transitory computer storage medium of claim 9 , wherein the query predicate is a range predicate, and wherein determining whether the particular in-memory data block satisfies the query predicate includes comparing a fingerprint associated with the particular in-memory data block and an encoding of the range predicate using a bitwise AND operation.

11. The non-transitory computer storage medium of claim 7 , wherein the determining and pruning are performed on the compressed columnar data and the generated fingerprints stored in the in-memory database.

12. The non-transitory computer storage medium of claim 7 , wherein any data block in the unpruned set of data blocks satisfies the query, and wherein decompressing the unpruned set of data blocks includes vectorizing values in the unpruned set of data blocks.

13. A computer-implemented system, comprising:

one or more computers; and

one or more computer memory devices interoperably coupled with the one or more computers and having tangible, non-transitory, machine-readable media storing one or more instructions that, when executed by the one or more computers, perform one or more operations comprising:

compressing columnar data to be stored in an in-memory database, including:

dividing the columnar data into a plurality of data blocks; and

in response to dividing the columnar data into the plurality of data blocks, for each particular data block in the plurality of data blocks, generating a particular fingerprint associated with the particular data block, wherein the particular fingerprint indicates a data range and one or more data gaps of the particular data block, and wherein the particular fingerprint is generated using one of a plurality of predefined fingerprint schemes based on data distribution in the columnar data;

storing the compressed columnar data and the generated fingerprints in the in-memory database;

receiving a query for the columnar data;

for each particular in-memory data block in the plurality of data blocks stored in the in-memory database:

determining whether the particular in-memory data block satisfies the query based on the particular fingerprint associated with the particular in-memory data block and on the data range and the one or more data gaps of the particular in-memory data block indicated by the particular fingerprint; and

in response to a determination that the particular in-memory data block does not satisfy the query, pruning the particular in-memory data block from the plurality of data blocks to generate an unpruned set of data blocks;

decompressing the unpruned set of data blocks; and

performing a query search on the decompressed unpruned set of data blocks for the received query.

14. The computer-implemented system of claim 5 , wherein the plurality of predefined fingerprint schemes includes at least two of a Basic Fingerprint (BF) scheme, a Unilateral Fingerprint (UF) scheme, a Selective Fingerprint (SF) scheme, a Two-Level Fingerprint (TLF) scheme, an Impartial Fingerprint (IF) scheme, and a Heterogeneous Fingerprint (HF) scheme.

15. The computer-implemented system of claim 13 , wherein the query includes a query predicate, and wherein determining whether the particular in-memory data block satisfies the query includes determining whether the particular in-memory data block satisfies the query predicate.

16. The computer-implemented system of claim 15 , wherein the query predicate is a range predicate, and wherein determining whether the particular in-memory data block satisfies the query predicate includes comparing a fingerprint associated with the particular in-memory data block and an encoding of the range predicate using a bitwise AND operation.

17. The computer-implemented system of claim 13 , wherein the determining and pruning are performed on the compressed columnar data and the generated fingerprints stored in the in-memory database.