System and a method for executing SQL basic operators on compressed data without decompression process
The present invention discloses a method for executing an SQL operator on compressed data chunk. The method comprising the step of: receiving SQL operator, accessing compressed data chunk blocks, receive e full set of derivatives of the compression scheme, check compression rules based on the compression scheme and relevant operator for approving SQL operation on compressed data and in case of approval applying respective SQL operator on relevant compressed data chunks.
1. A method for executing a Structure Query Language (SQL) operator on a compressed data chunk stored on a database management system (DBMS), said method comprising:
receiving the compression scheme of the compression data chunk, wherein the compression scheme is a Frame of Reference (FOR) scheme, wherein the compression scheme saves delta values based on offset values of each data chunk;
determining an algebraic transformation of the compression scheme;
responsively to the determination of the algebraic transformation, calculating numerical offset values of each compressed data chunk block of the compressed data chunk;
changing numerical offset values of all the compressed data chunk blocks to generate transformed compressed data chunk blocks having common offset values and recalculate the delta values accordingly, utilizing multiple threads of a Hardware Accelerator (HWA) unit to process the compressed data chunk blocks to generate the transformed compressed data chunk blocks, wherein the HWA has a superscalar vector architecture; and
applying the SQL operator having an algebraic operation to the transformed compressed data chunk blocks, utilizing the multiple threads of the HWA unit to apply the SQL operator to the transformed compressed data chunk blocks.
2. The method of claim 1 , wherein the SQL operator is applied to the compressed data chunk blocks only when the compression scheme includes an algebraic transformation.
3. The method of claim 2 , wherein the SQL operator is a sort operation.
4. The method of claim 2 , wherein the SQL operator is a merge operation and wherein utilizing multiple threads of the HWA unit comprises performing an offset alignment of all compressed data chunk blocks.
5. The method of claim 2 , wherein the SQL operator is join operation.
6. The method of claim 2 , wherein the SQL operator is a reduce operation having a commutative property.
7. The method of claim 1 , wherein the SQL operator is an injective hash function.
8. A system for executing a Structure Query Language (SQL) operator on a compressed data chunk, said system comprising:
a database configured to store a compressed data chunk including compression scheme;
at least one Hardware Accelerator (HWA) unit having a superscalar vector architecture;
at least one Central Processing Unit (CPU); and
at least one memory communicatively coupled to the at least one CPU comprising computer-readable instructions that when executed by the at least CPU cause the computer-based system to implement a method including receiving the compression scheme of the compressed data chunk, wherein the compression scheme is a Frame of Reference (FOR) scheme, wherein the compression scheme saves delta values based on offset values of each data chunk, determining an algebraic transformation of the compression scheme, responsively to the determination of the algebraic transformation, calculating numerical offset values of each compressed data chunk utilizing the compression scheme, applying the numeric offset values to the compressed data chunk to generate transformed compressed data chunk blocks having common offsets values and recalculate the delta values accordingly, utilizing multiple threads of a Hardware Accelerator (HWA) unit to process the compressed data chunk blocks to generate the transformed compressed data chunk blocks, wherein the HWA has a superscalar vector architecture; and applying the SQL operator having an algebraic operation to the transformed compressed data chunk blocks utilizing the multiple threads of the HWA unit to apply the SQL operator to the transformed compressed data chunk blocks.
9. The system of claim 8 , wherein the SQL operator is applied to the compressed data chunk blocks only when the compression scheme includes an algebraic transformation.
10. The system of claim 9 , wherein the SQL operator is a sort operation.
11. The system of claim 9 , wherein the SQL operator is a merge operation, and wherein utilizing multiple threads of the least one HWA unit comprises performing an offset alignment of the compressed data chunk blocks.
12. The system of claim 9 , wherein the SQL operator is a join operation.
13. The system of claim 9 , wherein the SQL operator is a reduce operation having a commutative property.
14. The system of claim 8 , wherein the SQL operator is an injective hash function.