Parallel processing system for traversing a transactional database

Claim 2

In a system including two or more processors, a computer implemented method for traversing one or more logical files, the logical files including two or more physical files of a transactional database, the method comprising the steps of: generating one or more sets of parameters for processing the two or more physical files of the transactional database; assigning a respective thread to each of the two or more physical files of the transactional database, each respective thread including instructions for processing its respective one of the two or more files of the transactional database in accordance with preselected ones of the one or more sets of processing parameters; sequentially traversing each of the two or more files of the transactional database once in accordance with its respective thread, at least two of the threads operating in parallel; retrieving, in accordance with the instructions for each thread, information from the respective one of the two or more physical files of the transactional database associated with said thread based upon the preselected ones of the one or more sets of processing parameters associated with said thread; storing the information retrieved by each thread in an output file.

Claim 3

The method according to claim 1 wherein the two or more files reside on one or more external mediums.

Claim 4

The method according to claim 2, further comprising the steps of: associating an input buffer set with each of the two or more files, each input buffer set including two or more input buffers; retrieving data blocks from each of the two or more files into respective ones of the two or more input buffers, in a same sequence as said data blocks are recorded on the physical medium.

Claim 5

The method according to claim 2, wherein the one or more external mediums include one or more of a tape drive, disk drive, a hard drive, an optical drive, a bubble memory and a flash memory.

Claim 6

In a system including two or more processors, a computer implemented method for traversing two or more physical files of a transactional database which correspond to a logical file based upon one or more view definitions, the method comprising the steps of: a) creating one or more view definitions, each view definition defining a set of processing parameters for selected ones of the two or more physical files of the transactional database; b) creating a logic table, the logic table including a logic table set for each of the two or more physical files of the transactional database; c) transforming the set of parameters for each view definition into a respective plurality of logic table entries, the plurality of logic table entries being copied into the logic table sets which correspond to the selected ones of the two or more physical files which correspond to the sets of parameters defined by each view definition; d) associating a respective one of a plurality of threads with each logic table set; e) transforming the logic table entries of each one of the plurality of threads, which correspond to each view definition associated with said each thread, into a set of instructions for said each thread; f) operating each thread in accordance with its set of instructions: to access each entry of its respective one of the two or more physical files of the transactional database, to retrieve information from said entry based upon the view definitions associated with said thread, and to store said retrieved information in one or more extract flies based upon the view definitions associated with said thread, at least two of the plurality of threads operating in parallel.

Claim 7

In a system including two or more processors, a computer implemented method for traversing two or more physical files of a transactional database which correspond to a logical file based upon one or more view definitions, the method comprising the steps of: a) creating one or more view definitions, each view definition defining a set of processing parameters for selected ones of the two or more physical files of the transactional database; b) creating a logic table, the logic table including a logic table set for each of the two or more physical files of the transactional database; c) transforming the set of parameters for each view definition into a respective plurality of logic table entries, the plurality of logic table entries being copied into the logic table sets which correspond to the selected ones of the two or more physical files which correspond to the sets of parameters defined by each view definition; d) associating a respective one of a plurality of threads with each logic table set; e) transforming the logic table entries of each one of the plurality of threads, which correspond to each view definition associated with said each thread, into a set of instructions for said each thread; f) operating each thread in accordance with its set of instructions: to access each entry of its respective one of the two or more physical files of the transactional database, to retrieve information from said entry based upon the view definitions associated with said thread, and to store said retrieved information in one or more extract flies based upon the view definitions associated with said thread, at least two of the plurality of threads operating in parallel.

Claim 8

The system according to claim 5, wherein the logic table is implemented as a "non-procedural" language.

Claim 9

The method according to claim 5, wherein the two or more files reside on one or more external mediums.

Claim 10

The method according to claim 7, wherein the one or more external mediums include one or more of a tape drive, a disk drive, a hard drive, an optical drive, a bubble memory and a flash memory.

Claim 11

The method according to claim 7, further comprising the steps of associating an input buffer set with each of the two or more files, each input buffer set including two or more input buffers; retrieving data blocks from each of the two or more files into respective ones of the two or more input buffers, in a same sequence as said data blocks are recorded on the physical medium.

Claim 12

The method according to claim 7, further comprising the steps of: a) associating an output buffer set with each of the one or more extract files, each output buffer set including one or more output buffers; and wherein the step of operating each thread to store said retrieved information in one or more extract files further includes the steps of b) storing said information retrieved by said each thread in a current one of the one or more output buffers associated with a one of the one or more extract files; c) determining whether said current one of the one or more output buffers is full, and, when said current one of the one or more output buffers is determined to be full, 1) initiating a transfer of the current one of the one or more output buffers to its associated extract file, and 2) designating another one of the output buffers as the current one of the one or more output buffers.

Claim 13

A parallel processing system for traversing a logical file of a transactional database, the logical file including two or more physical files, the parallel processing system comprising: an input buffer for retrieving blocks of data from the two or more physical files of the transactional database; an output buffer for storing blocks of data in one or more extract files; a logic table storing one or more sets of processing parameters, each set of processing parameters corresponding to a respective view definition, each view definition defining criteria for retrieving information from one or more files of the logical record, each view definition further defining a format of a designated one of the extract files, the logic table including at least two logic table sets, each logic table set associated with a respective file of the two or more files of the transactional database, each logic table set including the respective set of processing parameters associated with each view definition which defines criteria for retrieving information from the respective file associated with said logic table set; a corresponding thread associated with each logic table set, each thread implementing the sets of processing parameters associated with its respective logic table set, each thread sequentially accessing entries from the file of the transactional database associated with its respective logic table set, each thread applying each of its sets of processing parameters to each accessed entry such that each entry is accessed only once, each thread retrieving information from each accessed entry in accordance with its sets of processing parameters and storing the retrieved information in the designated extract files, via the output buffer, in the respective formats defined by the view definitions corresponding to the sets of processing parameters.

Claim 14

A parallel processing system for traversing a logical file of a transactional database, the logical file including two or more physical files, the parallel processing system comprising: an input buffer for retrieving blocks of data from the two or more physical files of the transactional database; an output buffer for storing blocks of data in one or more extract files; a logic table storing one or more sets of processing parameters, each set of processing parameters corresponding to a respective view definition, each view definition defining criteria for retrieving information from one or more files of the logical record, each view definition further defining a format of a designated one of the extract files, the logic table including at least two logic table sets, each logic table set associated with a respective file of the two or more files of the transactional database, each logic table set including the respective set of processing parameters associated with each view definition which defines criteria for retrieving information from the respective file associated with said logic table set; a corresponding thread associated with each logic table set, each thread implementing the sets of processing parameters associated with its respective logic table set, each thread sequentially accessing entries from the file of the transactional database associated with its respective logic table set, each thread applying each of its sets of processing parameters to each accessed entry such that each entry is accessed only once, each thread retrieving information from each accessed entry in accordance with its sets of processing parameters and storing the retrieved information in the designated extract files, via the output buffer, in the respective formats defined by the view definitions corresponding to the sets of processing parameters.

Claim 15

The system according to claim 11, wherein the logic table is implemented as a "non-procedural" language.

Claim 16

The system according to claim 11, wherein the input buffer includes a corresponding input buffer set for each thread, each corresponding input buffer set including one or more input buffers, each input buffer for storing a retrieved data block, a size of the data block being selected as a function of a physical medium from which the data block was retrieved.

Claim 17

The system according to claim 11, wherein the output buffer includes a corresponding output buffer set for each extract file, each corresponding output buffer including one or more output buffers, each output buffer for holding a data block for storage in the extract file associated with the corresponding output buffer, a size of each data block being selected as a function of a physical medium in which the extract file resides.

Claim 18

The system according to claim 13, wherein the thread associated with the one or more input buffers of each input buffer set implements its sets of processing parameters on a current one of the one or more input buffers while another processor retrieves a next data block from the one of the two or more files associated with the thread.

Claim 19

The system according to claim 14, wherein each thread, prior to storing the retrieved information in the output buffer set corresponding to the designated extract file, determines whether a current output buffer of the output buffer set corresponding to the designated extract file is full, and, if the current output buffer is not full, stores the retrieved information in the current output buffer, and if the current output buffer is full, initiates a transfer of the current one of the one or more output buffers to the designated extract file and designates another one of the output buffers as the current one of the one or more output buffers.

Claim 20

The system according to claim 13, wherein each input buffer set includes a plurality of input buffers.

Claim 21

The system according to claim 14, wherein each output buffer set includes a plurality of output buffers.

Claim 22

The system according to claim 13, wherein the physical medium includes one of a tape drive, a disk drive, a hard drive, an optical drive, a bubble memory and a flash memory.

Claim 23

The system according to claim 14, wherein the physical medium includes one of a tape drive, a disk drive, a hard drive, an optical drive, a bubble memory and a flash memory.