Unified architecture for hybrid database storage using fragments

1. A computer program product comprising a non-transitory machine-readable medium storing instructions that, when executed by at least one programmable processor, cause the at least one programmable processor to perform operations comprising:

storing, in a plurality of fragments retained in on-disk storage, a plurality of data records that comprise a data set, each fragment of the plurality of fragments comprising a local row index, a mapping of the local row index to RowID values in a global RowID column for the data set, and a number of data records that is equal to or less than a defined maximum fragment size, the RowID values in the RowID column being unique, logical references to the data records of the data set, the local row index comprising physical locations of the data records in the fragment, the defined maximum fragment size being smaller than an available amount of main system memory to facilitate the loading of each fragment of the plurality of fragments into the main system memory;

compressing each fragment of the plurality of fragments using a compression that is optimized for each fragment;

reading at least one of the plurality of fragments into main system memory from the on-disk storage; and

performing an operation on the at least one of the plurality of fragments while the at least one of the plurality of fragments is in the main system memory;

executing a merge of the data set with a change to the data set, the merge comprising:

creating a temporary, in-memory data structure containing data from a fragment of the plurality of fragments and a committed row from the change to the data set, and

sorting the temporary, in-memory data structure to maximize compression of the plurality of data records; and

storing data from the sorted, temporary, in-memory data structure to a revised fragment.

2. The computer program product as in claim 1 , wherein each of the plurality of part fragments comprises at least one row set arranged in a temporal order.

3. The computer program product as in claim 1 , wherein the merge further comprises:

copying one or more uncommitted rows from the change of the data set to a delta part of a record.

4. The computer program product as in claim 3 , wherein the operations further comprise: storing cleanup pointers at a start of the merge, the cleanup pointers indicating start and stop locations for the temporary data structure.

5. A system comprising:

at least one processor; and

a machine-readable medium storing instructions that, when executed by the at least one programmable processor, cause the at least one programmable processor to perform operations comprising:

storing, in a plurality of fragments retained in on-disk storage, a plurality of data records that comprise a data set, each fragment of the plurality of fragments comprising a local row index, a mapping of the local row index to RowID values in a global RowID column for the data set, and a number of data records that is equal to or less than a defined maximum fragment size, the RowID values in the RowID column being unique, logical references to the data records of the data set, the local row index comprising physical locations of the data records in the fragment, the defined maximum fragment size being smaller than an available amount of main system memory to facilitate the loading of each fragment of the plurality of fragments into the main system memory;

compressing each fragment of the plurality of fragments using a compression that is optimized for each fragment;

reading at least one of the plurality of fragments into main system memory from the on-disk storage; and

performing an operation on the at least one of the plurality of fragments while the at least one of the plurality of fragments is in the main system memory;

executing a merge of the data set with a change to the data set, the merge comprising:

creating a temporary, in-memory data structure containing data from a fragment of the plurality of fragments and a committed row from the change to the data set, and

sorting the temporary, in-memory data structure to maximize compression of the plurality of data records; and

storing data from the sorted, temporary, in-memory data structure to a revised fragment.

6. The system as in claim 5 , wherein each of the plurality of part fragments comprises at least one row set arranged in a temporal order.

7. The system as in claim 5 , wherein the merge further comprises:

copying one or more uncommitted rows from the change of the data set to a delta part of a record.

8. The system as in claim 7 , wherein the operations further comprise: storing cleanup pointers at a start of the merge, the cleanup pointers indicating start and stop locations for the temporary data structure.

9. A computer-implemented method comprising:

storing, in a plurality of fragments retained in on-disk storage, a plurality of data records that comprise a data set, each fragment of the plurality of fragments comprising a local row index, a mapping of the local row index to RowID values in a global RowID column for the data set, and a number of data records that is equal to or less than a defined maximum fragment size, the RowID values in the RowID column being unique, logical references to the data records of the data set, the local row index comprising physical locations of the data records in the fragment, the defined maximum fragment size being smaller than an available amount of main system memory to facilitate the loading of each fragment of the plurality of fragments into the main system memory;

compressing each fragment of the plurality of fragments using a compression that is optimized for each fragment;

reading at least one of the plurality of fragments into main system memory from the on-disk storage; and

performing an operation on the at least one of the plurality of fragments while the at least one of the plurality of fragments is in the main system memory;

executing a merge of the data set with a change to the data set, the merge comprising:

creating a temporary, in-memory data structure containing data from a fragment of the plurality of fragments and a committed row from the change to the data set, and

sorting the temporary, in-memory data structure to maximize compression of the plurality of data records; and

storing data from the sorted, temporary, in-memory data structure to a revised fragment.

10. The computer-implemented method as in claim 9 , wherein each of the plurality of part fragments comprises at least one row set arranged in a temporal order.

11. The computer-implemented method as in claim 9 , wherein the merge further comprises:

copying one or more uncommitted rows from the change of the data set to a delta part of a record.

12. The computer-implemented method as in claim 11 , further comprising: storing cleanup pointers at a start of the merge, the cleanup pointers indicating start and stop locations for the temporary data structure.

13. The computer-implemented method as in claim 9 , wherein at least one of the storing, the compressing, the reading, the performing, and the executing are performed by a system comprising at least one programmable processor.