Each datasource has one table, including at least the following fields:
compound identifier (Primary CID)
compound structure (InChi, canonical SMILES)
target identifier ( Uniprot or gene name or entrez gene ID or EC number)

Optional:
compound name, casNO, compound properties
target name, sequence info,
interaction type
binding affinity

One table summerizes the provenance of all the datasource, including:
datasource name
corresponding table name
the link of datasource
download date
who