When writing to a shift-sensitive flat file target, the Integration Service adds shift characters and spaces if the data going into the target does not meet file requirements. You need to allow at least two extra bytes in each data column containing multibyte data so the output data precision matches the byte width of the target column.

The Integration Service writes shift characters and spaces in the following ways:

To illustrate how the Integration Service handles a fixed-width file containing shift-sensitive data, say you want to output the following data to the target:

A

is a double-byte character,

a

is a single-byte character.

The first target column contains eight bytes and the second target column contains four bytes.

The Integration Service must add shift characters to handle shift-sensitive data. Since the first target column can handle eight bytes, the Integration Service truncates the data before it can add the shift characters.

The following table describes the notation used in this example:

For the first target column, the Integration Service writes three of the double-byte characters to the target. It cannot write any additional double-byte characters to the output column because the column must end in a single-byte character. If you add two more bytes to the first target column definition, then the Integration Service can add shift characters and write all the data without truncation.

For the second target column, the Integration Service writes all four single-byte characters to the target. It does not add write shift characters to the column because the column begins and ends with single-byte characters.