Identifying joins of tables of a database

A device for identifying table joins, comprising:

• a memory;
• and
• a processor, the processor configured to execute instructions stored in the memory to:
• obtain respective casting similarities between pairs of columns of a first table and a second table, wherein a pair of the pairs of columns comprises a first column of the first table and a second column of the second table, and wherein a casting similarity obtained for the pair of columns is based on an extent to which first data values of the first column are changeable to a data type of the second column;
• discard, to obtain first join candidates, ones of the pairs of columns not satisfying a casting similarity condition;
• obtain, based on respective edit distances, respective string similarities for the first join candidates, wherein an edit distance between a first column name of a column of the first table and a second column name of a column of the second table indicates a number of operations required to transform the first column name into the second column name;
• discard ones of the first join candidates not satisfying a string similarity condition to obtain second join candidates;
• obtain final join candidates using the respective casting similarities and the respective string similarities of the second join candidates;
• present the final join candidate on a device of a user;
• receive, from the device of the user, a selected join candidate of the final join candidates;
• and
• query a database based on a data query that includes a join of the first table and the second table to obtain tabular data, wherein the join is based on the selected join candidate.

A non-transitory computer-readable storage medium, comprising executable instructions that, when executed by a processor, facilitate performance of operations for identifying table joins, the operations comprising:

• obtaining respective casting similarities between pairs of columns of a first table and a second table, wherein a pair of columns of the pairs of columns comprises a first column of the first table and a second column of the second table, and wherein a casting similarity for the pair of columns is obtained by steps comprising:
• assigning the casting similarity to the pair of columns based on an identified extent to which first data values of the first column are changeable to a data type of the second column, wherein the casting similarity is selected from a set comprising a very low casting similarity, and wherein the very low casting similarity is assigned to a given pair of columns in a case that one column of the given pair of columns has a BOOLEAN type and the other column of the given pair of columns has a FLOAT type;
• discarding, to obtain first join candidates, ones of the pairs of columns having the respective casting similarities not satisfying a casting similarity threshold;
• obtaining respective string similarities for the first join candidates;
• discarding ones of the first join candidates not satisfying a string similarity condition to obtain second join candidates;
• obtaining final join candidates using the respective casting similarities and the respective string similarities of the second join candidates, wherein each of the final join candidates includes a column of the first table and a column of the second table;
• presenting the final join candidate on a device of a user;
• receiving, from the device of the user, a selected join candidate of the final join candidates;
• querying a database based on a data query that includes a join of the first table and the second table to obtain tabular data, wherein the join is based on the selected join candidate;
• and
• outputting the tabular data.