Modeling Guide

Format Converter

The Format Converter operator converts blobs between CSV, JSON and XML. The output format is set as a configuration value. The input format is deduced using the following rules:

The input format is deduced using the following rules:
  • If it starts with { or [, it must be JSON.

  • If it starts with <, it must be XML.

  • Otherwise, it must be CSV.

If the first field in CSV input starts with any of the three special characters ({, [ and <), it must be quoted.

CSV

The CSV format conforms to RFC 4180Information published on non-SAP site, with the following observations:
  • any valid UTF-8 character or escape sequence may be used as field separator ("comma").

  • output CSV will use CRLF as a record separator, placing it after every record including the last one.

  • input CSV can use either LF or CRLF as a record separator, and the last record may or may not be followed by a separator.

  • since CSV lacks any type information, when converting to JSON, all fields will be made into strings.

XML

Converting to XML requires one of the following to be true:
  • xmlOutputRootTag is not empty, in which case all input content will be placed within this tag. For example, input
    [{"a": 1}, {"b": "hello"}]
    becomes
    <root>
        <item>
            <a>1</a>
        </item>
        <item>
            <b>hello</b>
        </item>
    </root>
    where xmlOutputRootTag is root and xmlOutputItemTag is item.
  • Input is a JSON object containing a single property, whose key will be used as root tag. For example:
    {"countries": [{"id": "US", "currency": "USD"}, {"id": "DE", "currency": "EUR"}]}
    becomes
    <countries>
        <country>
            <id>US</id>
            <currency>USD</currency>
        </country>
        <country>
            <id>DE</id>
            <currency>EUR</currency>
        </country>
    </countries>
    assuming xmlOutputRootTag is empty and xmlOutputItemTag is country.
Converting from XML requires no XML-specific configuration. Sibling tags with the same name will be placed in an array named after them. For example:
<people>
    <person>
        <name>Bonnie</name>
        <age>23</age>
    </person>
    <person>
        <name>Clyde</name>
        <age>25</age>
    </person>
</people>
will be turned into
{
    "people": {
        "person": [
            {"name": "Bonnie", "age": "23"},
            {"name": "Clyde", "age": "25"}
        ]
    }
}
Notice that all textual values in XML become JSON strings.

Configuration Parameters

Parameter

Type

Description

csvComma

string

The field separator.

Default: ","

targetFormat

string The file type to be converted to. Accepted values are JSON, CSV or XML.

Default: "JSON"

csvHeaderIncluded

bool

Indicates whether or not each CSV input will have a header as first line. If the header will not be present, the fields parameter must be supplied.

Default: false

fields

string

A comma-separated list of field names. This value is only required when:
  • converting from headerless CSV

  • converting to CSV

Default: ""

xmlOutputRootTag

string

The name for the root tag to be used when converting to XML. If this value is empty, the operator will only be able to convert JSON input, which must be an object with a single property that will serve as root tag.

Default: "items"

xmlOutputItemTag

string

The tag name to be used for each element in an array when converting to XML. If this value is empty, xmlOutputRootTag will be used.

Default: "item"

Input

Input

Type

Description

input

blob

The source data to be converted.

Output

Output

Type

Description

output

blob

The resulting data, in targetFormat format.