An input file without tabular structure in which every line contains its data elements one after the other separated by a distinguishing character (e.g., a semicolon) must be loaded with the COMPRESS option. This option runs:
In COMPRESS mode, the position specifications after the column names denote the relative position of the values in the file. To find a value designated by <n>, <n> - 1 separators defined by the SEPARATOR specification are skipped when reading. Separators may also follow immediately one after the other (input length 0). If the relative position is greater than the number of values in the line, the value, too, has the input length 0. By definition, values of length 0 satisfy the NULL condition of the column.
The DELIMITER specification serves to delimit character strings, for example, when these contain the separator symbol. Exactly those characters will be inserted which are placed between the two delimiters. If the first non-empty character of a value is not the delimiter, then all characters up to the next separator will be inserted.
Specifying COMPRESS is equivalent to the specification of SEPARATOR ',' and DELIMITER '"'.