Catalog data files

Essentially a two-dimensional table, a catalog data file consists of a header record which identifies the data columns and any number of data records (rows). Fields in both header and data records are separated by single <TAB> characters. Records are separated by a single <CR> (ASCII code 0xD), a single <LF> (ASCII code 0xA), or a <CR><LF> pair.

The header record must contain the exact names for each data field. Empty fields are not permitted in the header row. Data field names are not case-sensitive. All field names must be unique.

No white space other than the <TAB> field separators is permitted in header and data records.

In the data records, two adjacent <TAB> characters indicate an empty field. Empty fields inherit default values from either the catalog attributes or from the server defaults.

Data fields must not contain <CR>, <LF>, or <TAB> characters, unless the data value is of type text and is enclosed by single or double quotation marks. Do not HTTP-encode data fields.

Multiple data values in the same field are separated by commas (‘,’), unless otherwise noted.

Columns whose names start with ‘.’ are ignored; this permits storing data in material catalogs which is of no interest to Image Rendering. Columns with unknown header names are ignored and a warning is written to the log file.

Field names can consist of any combination of ASCII letters, numbers, and “-” and “_”.

One or more columns can be used as index keys. If the same key occurs more than once in the same data file, the later instance prevails.