Heuristic dimension reduction in metadata modeling

A computer program product comprising:

• a computer readable storage medium readable by one or more processor and storing instructions for execution by the one or more processor for performing a method comprising:
• obtaining metadata on respective datasets from a plurality of data sources;
• preparing an enhanced metadata based on the metadata from the obtaining, wherein an item of the metadata is associated with semantic information for the item, available data statistics on the item, and a search index for the item, and wherein the enhanced metadata has all items of the metadata semantically classified into respective potential join columns;
• receiving an intent phrase for search on the datasets from the plurality of data sources;
• and
• producing a metadata model that satisfies the intent phrase, the metadata model including a plurality of join columns and a plurality of join relationships, wherein a join column of the metadata model indicates a column common in two entries from the enhanced metadata, wherein a join relationship of the metadata model is inferred as a relationship between two entries of the enhanced metadata that commonly have the join column, and wherein the two entries are classified into respective semantic element classes extracted from the intent phrase.

A system comprising:

• a memory;
• one or more processor in communication with memory;
• and
• program instructions executable by the one or more processor via the memory to perform a method comprising:
• obtaining metadata on respective datasets from a plurality of data sources;
• preparing an enhanced metadata based on the metadata from the obtaining, wherein an item of the metadata is associated with semantic information for the item, available data statistics on the item, and a search index for the item, and wherein the enhanced metadata has all items of the metadata semantically classified into respective potential join columns;
• receiving an intent phrase for search on the datasets from the plurality of data sources;
• and
• producing a metadata model that satisfies the intent phrase, the metadata model including a plurality of join columns and a plurality of join relationships, wherein a join column of the metadata model indicates a column common in two entries from the enhanced metadata, wherein a join relationship of the metadata model is inferred as a relationship between two entries of the enhanced metadata that commonly have the join column, and wherein the two entries are classified into respective semantic element classes extracted from the intent phrase.