Index and data blocks consist of a block header followed by a series of records.

The block header has three fields that contain information. The first field, of two bytes, specifies the number of bytes currently in use in the block. The second field, of one byte, specifies the block level. The last field of four bytes represents the transaction number at which the block was last changed.

Depending on the platform, there may also be an empty field containing filler to produce proper alignment. The filler occurs between the second and third data field and causes the length of the header to increase from seven to eight bytes. You can determine the size of the header by dumping the data block with DSE.