Data formats determine how data is written when transferred between systems. The most common are the delimited tabular format, which is simple and readable but cannot handle nested structures, the format used for interfaces, which does support nested data, and an older markup format widespread in industrial and accounting systems. Most practical problems do not arise from the choice of format but from character encoding and delimiters – scrambled diacritics and shifted columns are almost always the result of incorrect encoding or a comma within text. When agreeing terms with a supplier, it is therefore worth specifying, besides the format, also the encoding, delimiter, date and decimal number format, and how empty values are to be written. A sample file with a few rows agreed by both sides saves more time during integration than any written specification.
See also: Product feed, Data normalisation, API as a data source.
