Synthesizing mapping relationships using table corpus

1. A system for synthesizing mapping relationships using a table corpus, comprising:

a processor; and

a memory storage device including instructions that when executed by the processor are operable to:

extract a table from a corpus of tables, the table including a plurality of columns;

remove a column comprising incoherent values from the plurality of columns;

generate column pairs from remaining columns, the generated column pairs having a binary relationship between a first column of the generated column pair and a second column of the generated column pair;

remove each of the generated column pairs that has a binary relationship lacking a functional dependency; and

synthesize remaining generated column pairs to determine at least one mapping table, the at least one mapping table comprising a first column having the functional dependency with a second column.

2. The system of claim 1 , wherein to remove the column comprising the incoherent values from the plurality of columns, the instructions that when executed by the processor are further operable to:

determine a coherence score for each column of the plurality of columns; and

remove the column having the coherence score less than a predetermined threshold from the plurality of columns.

3. The system of claim 1 , wherein to synthesize the remaining generated column pairs, the instructions that when executed by the processor are further operable to:

determine a compatibility score between each column pair of the remaining generated column pairs; and

merge a first column pair of the remaining generated column pairs based on the determined compatibility score.

4. The system of claim 3 , wherein the instructions that when executed by the processor are operable to determine a positive compatibility score and a negative compatibility score between each column pair of the remaining generated column pairs.

5. A method for synthesizing mapping relationships using a table corpus, comprising:

extracting a table from a corpus of tables, the table including a plurality of columns;

removing a column comprising incoherent values from the plurality of columns;

generating two-column tables from remaining columns, the generated two-column tables having a binary relationship between a first column of the two-column table and a second column of the two-column table;

removing each of the generated two-column tables that has a binary relationship lacking a functional dependency; and

synthesizing, based on remaining generated two-column tables, at least one mapping table, the at least one mapping table comprising a first column having the functional dependency with a second column.

6. The method of claim 5 , wherein removing the column comprising the incoherent values from the plurality of columns comprises:

determining a coherence score for each column of the plurality of columns; and

removing the column from the plurality of columns when the coherence score for the column is less than a predetermined threshold.

7. The method of claim 6 , wherein the coherence score for each column is determined based on a frequency of occurrence of each item in the column.

8. The method of claim 5 , wherein synthesizing the at least one mapping table comprises determining a compatibility score for pairs of the remaining generated two-column tables.

9. The method of claim 8 , wherein determining the compatibility score comprises determining a positive compatibility score comprising an overlap between items of the pairs of the remaining generated two-column tables.

10. The method of claim 9 , wherein determining the positive compatibility score further comprises determining the overlap between the items of the pairs of the remaining generated two-column tables based on approximate string matching between the items of the pairs.

11. The method of claim 9 , further determining a negative compatibility score comprising conflicting items between the pairs of the remaining generated two-column tables.

12. A computer readable storage device including computer readable instructions for synthesizing mapping relationships using a table corpus, which when executed by a processing unit is operative to:

determine a functional dependency between at least two items of an input table;

extract a table from a corpus of tables, the table including a plurality of columns;

remove a column comprising incoherent values from the plurality of columns;

generate two-column tables from remaining columns, the generated two-column tables having a binary relationship between a first column of the two-column table and a second column of the two-column table;

remove each of the generated two-column tables that has a binary relationship lacking the functional dependency; and

synthesize remaining generated two-column tables to determine at least one mapping table, the at least one mapping table comprising a first column having the functional dependency with a second column; and

provide a next item of the input table from the determined at least one mapping table.

13. The computer readable storage device of claim 12 , wherein to remove the column comprising the incoherent values from the plurality of columns, the computer readable instructions when executed by the processing unit is further operative to:

determine a coherence score for each column of the plurality of columns; and

remove the column having the coherence score below a predetermined threshold from the plurality of columns.

14. The computer readable storage device of claim 12 , wherein to synthesize the remaining generated two-column tables the computer readable instructions when executed by the processing unit is further operative to:

determine a compatibility score between pairs of the remaining generated two-column tables; and

merge a first pair of the remaining generated two-column tables based on the determined compatibility score.

15. The computer readable storage device of claim 14 , wherein the computer readable instructions when executed by the processing unit is further operative to:

determine the compatibility score using approximate string matching between values in each column of the pairs of the remaining generated two-column tables.