Index files are auxiliary files used to speed up the searching of a data file. The records in an index file have only two fields, namely, the key-value and the address of a page in the data file.
Records in an index file are often ordered (possibly using a space filling curve) and may be further organised using specialised search data structures such as B-tree, R-tree. Grid file, and so on. Figure 5.1 shows a secondary index file of city table on the key field name.
Because the records in index files are only thirty eight bytes, assuming eight byte addresses for disk pages, all nine index records fit in a single disk page.
The index records are sorted and contain addresses of disk pages in the data file containing appropriate records. The data file itself may be unordered with respect to the key field.
