Data matching includes receiving a piece of first relational data and a piece of second relational data. The piece of first relational data is associated with a plurality of pieces of first data, and the piece of second relational data is associated with a plurality of pieces of second data. An approximate value of the piece of second relational data is calculated. A similarity is calculated based on the piece of first relational data and the approximate value of the piece of second relational data. A correspondence between a piece of the first data and a piece of the second data is determined based on the calculated similarity. An alignment parameter is calculated based on the determined correspondence, and a first data group including the piece of the first data is matched with a second data group including the piece of the second data based on the alignment parameter.