Concurrent optimistic transactions for tables with deletion vectors

A system comprising:

• one or more computer processors;
• and
• one or more computer-readable mediums storing instructions that, when executed by the one or more computer processors, cause the system to perform operations comprising:
• receiving, from a client device, a request to read records in a data table, the data table being associated with a deletion vector, wherein the deletion vector is a metadata structure stored in association with the data table, and wherein each value in the deletion vector corresponds to a relative one of the records in the data table;
• identifying a first value in the deletion vector corresponding to a first record in the data table;
• determining that the first value indicates that the first record has been deleted based on a previously executed transaction to the data table;
• returning, to the client device, a subset of the records in the data table based on determining that the first value indicates that the first record has been deleted, the subset of records not including the first record;
• receiving a concurrent request to generate an updated version of the data table, the updated version of the data table including a modification to at least one existing record in the data table;
• determining the at least one existing record for the updated version of the data table can be located in at least a subset of data files of the data table;
• responsive to determining that the records can be located, updating a value in the deletion vector corresponding to the at least one existing record in the data table to indicate that the at least one existing record is deleted;
• and
• committing the updated version of the data table to a metadata log file.

A non-transitory computer-readable medium storing instructions that, when executed by one or more computer processors of a computing system, cause the computing system to perform operations comprising:

• receiving, from a client device, a request to read records in a data table, the data table being associated with a deletion vector, wherein the deletion vector is a metadata structure stored in association with the data table, and wherein each value in the deletion vector corresponds to a relative one of the records in the data table;
• identifying a first value in the deletion vector corresponding to a first record in the data table;
• determining that the first value indicates that the first record has been deleted based on a previously executed transaction to the data table;
• returning, to the client device, a subset of the records in the data table based on determining that the first value indicates that the first record has been deleted, the subset of records not including the first record;
• receiving a concurrent request to generate an updated version of the data table, the updated version of the data table including a modification to at least one existing record in the data table;
• determining the at least one existing record for the updated version of the data table can be located in at least a subset of data files of the data table;
• responsive to determining that the records can be located, updating a value in the deletion vector corresponding to the at least one existing record in the data table to indicate that the at least one existing record is deleted;
• and
• committing the updated version of the data table to a metadata log file.