Encoding a spreadsheet program for stream processing

A system for encoding a function in a spreadsheet program to support stream processing, the system comprising:

• at least one computer system including at least one processor unit and a memory connected to the at least one processor unit;
• said at least one processor unit including:
• at least one encoding module for encoding one or more formulas in one or more cells of the spreadsheet program, the one or more formulas supporting data windows and stream partitioning in a data stream input to the spreadsheet program;
• and
• at least one spreadsheet module for designating a first group of cells of the spreadsheet program as input cells for holding input values for said one or more formulas, and for designating a second group of cells of the spreadsheet program as output cells for holding output values of said one or more formulas;
• and wherein:
• one or more of the data windows in the data stream input to the spreadsheet program includes a plurality of stream values;
• the one or more formulas supporting data windows and stream partitioning includes
• supporting data windows by encoding one of the cells of the spreadsheet with one formula of said one or more formulas to operate on all the values of each of said one or more of the data windows directly from the each data window in the data stream input to the spreadsheet program into an aggregate value for the one of the cells to decouple the data windows in the data stream input to the spreadsheet program from the spreadsheet, and
• supporting stream partitioning by encoding the spreadsheet program to partition the input data stream into a plurality of data stream partitions, and wherein the one or more formulas encoded in the spreadsheet program operate on each of the data stream partitions;
• and
• the at least one spreadsheet module includes
• a common class for implementing a windowing mechanism for inserting new ones of the stream values and evicting old ones of the stream values, and
• a specialized class for implementing a logic for said one formula;
• and wherein:
• each of the stream values in the data stream has an associated timestamp;
• and
• the designating a first group of cells of the spreadsheet program as input cells for holding input values for said one or more formulas includes designating one of the first group of cells to hold one of the input values to represent the stream values in the input stream having timestamps within a specified period of time;
• and
• the supporting data windows further includes providing code for
• connecting an operator to the input data stream,
• naming a file containing the spreadsheet program,
• assigning attributes of input triples of the input data stream to cells of the first group of cells of the spreadsheet program,
• identifying ones of the attributes serving as partitioning keys, and
• identifying ones of the attributes serving as the timestamps.

A computer program product comprising:

• a computer readable medium having computer program code tangibly embodied therein for encoding a function in a spreadsheet program to support stream processing, the computer program code, when executed in a computer system, performing the following:
• encoding one or more formulas in one or more cells of the spreadsheet program, each of the formulas supporting data windows and stream partitioning in a data stream input to the spreadsheet program;
• designating a first group of cells of the spreadsheet program as input cells for holding input values for said one or more formulas;
• and
• designating a second group of cells of the spreadsheet program as output cells for holding output values of said one or more formulas;
• and wherein:
• one or more of the data windows in the data stream input to the spreadsheet program includes a plurality of stream values;
• and
• the one or more formulas supporting data windows and stream partitioning includes
• supporting data windows by encoding one of the cells of the spreadsheet with one formula of said one or more formulas to operate on all the stream values of each of said one or more of the data windows directly from the each data window in the data stream input to the spreadsheet program into an aggregate value for the one of the cells to decouple the data windows in the data stream input to the spreadsheet program from the spreadsheet, including
• providing a common class for implementing a windowing mechanism for inserting new ones of the stream values and evicting old ones of the stream values, and
• providing a specialized class for implementing a logic for said one formula,
• and
• supporting stream partitioning by encoding the spreadsheet program to partition the input data stream into a plurality of data stream partitions, and wherein the one or more formulas encoded in the spreadsheet program operate on each of the data stream partitions;
• and wherein:
• each of the stream values in the data stream has an associated timestamp;
• and
• the designating a first group of cells of the spreadsheet program as input cells for holding input values for said one or more formulas includes designating one of the first group of cells to hold one of the input values to represent the stream values in the input stream having timestamps within a specified period of time;
• and
• the supporting data windows further includes providing code for
• connecting an operator to the input data stream,
• naming a file containing the spreadsheet program,
• assigning attributes of input triples of the input data stream to cells of the first group of cells of the spreadsheet program,
• identifying ones of the attributes serving as partitioning keys, and
• identifying ones of the attributes serving as the timestamps.