An image data combining apparatus combines m lines of image data that have been output from a reading unit. The reading unit outputs m lines of image data based upon the pixel data that has been output upon being divided into the plurality of lines, and stored then in a memory. An upper address for accessing the memory is decided based upon first data indicating position, along the sub-scan direction, of an image that has been read by the reading unit, and a lower address for accessing the memory is decided based upon second data indicating position of the image along the main-scan direction. For the lower address utilizes values in which the sequence of a plurality of bits constituting the first data is interchanged, so that p items of pixel data at a time are extracted successively from each of the m lines of image data.