The database must be converted to the correct format before it can be used by the control system. The format conversion is done by a family of programs that all have roughly the same structure.

The table to be converted is extracted from the Postgres database by a SQL query and placed in a file. When the table is extracted, a filter is applied to substitute integer values for names on certain fields. For example, in the DescRec table, the name in the DataType field is substituted for the integer value known by the runtime system. In the above example, the Dkeys table is considered an immutable table. This means that it cannot/should not be changed. The table is in place to prevent bad entries in database. The actual integer values are defined in the filter/AccelNET source code. The Dkeys table will only be changed when there is a change in AccelNET that affects it. The result of this is an ASCII file where each line contains a number of fields separated by a field separator (usually a vertical bar '|') and terminated by a linefeed.

A program then translates the created file into the record structure used by the control system. This process involves steps such as converting strings of characters into floating point numbers. The program that does the translation is specifically tailored for the table in question, producing a new result file.

It is presently arranged so that the extracted Postgres tables occupy a directory. The SQL scripts share the directory with the extracted files along with a shell script that is used to translate the tables.

The table translation programs, invoked from a shell script, are in a separate directory. They read from the directory containing the extracted files and write into another directory.