You can configure Comma-Separated Value (CSV) parsers for both FileLog and WinLog collectors.

The available options for the csv parser are fields and delimiter.

Comma-Separated Value Parser Options

Note the following information about the structure of the csv parser.




The fields option specifies the names of the fields that exist in the log. The total number of the listed field names must be equal to the total number of comma-separated fields in the logs.

The fields option is mandatory for the CSV parser. If it is not specified, nothing is parsed. Double quotes surrounding the field value are optional, depending on the field content.

Field names must be separated by commas, for example

fields = field_name1, field_name2, field_name3, field_name4

This definition assumes that the names field_name1, field_name2, field_name3 and field_name4 are assigned sequentially to the extracted fields.

If some fields must be omitted by the CSV parser, their names can be omitted from the list. For example,

 fields = field_name1, , field_name3, field_name4

In this case, the parser extracts only the first, third and fourth fields from the event and subsequently assigns the names field_name1, field_name3 and field_name4 to them.

If the fields option does not specify a complete list of the fields in your logs, the parser returns an empty list. For example, if the log file contains field1, field2, field3, field4, and field5, but only fields= field1,field2,field3 is specified, the parser returns an empty fields list.

You cannot use fields=* for a CSV parser, because the parser returns an empty fields list. A complete list of fields must be specified, unless you need certain fields omitted as already described.


The delimiter option specifies the delimiter for the parser to use. By default, the csv parser uses a comma as a delimiter; however, you can change the delimiter to a semicolon, a space, or other special character. The defined delimiter must be enclosed in double quotes.

For example, delimiter="," and delimiter=";".

The csv parser supports any set of characters as delimiters that are enclosed in quotes, for example "||" or "asd". The field values' separators in the logs should exactly match the pattern defined by the delimiter parameter, otherwise the parser will fail.

Special characters such as a space or a tab can be defined for as a delimiter for the csv parser as long as the escape character precedes the special character for (\", \s, \t). For example, delimiter="\s" or delimiter=" ".

The delimiter option is optional.

CSV Log Parser Configuration

To parse logs collected from either winlog or filelog sources, use the following configuration.


base_parser = csv
fields = timepstamp,field_name1, field_name2, field_name3
delimiter = ";"
field_decoder={"timestamp": "tsp_parser"}
; timestamp is a built-in parser
; "format" is an option of timestamp parser 
format=%Y-%m-%d %H:%M:%S

With this configuration, logs collected from some_csv_logs source (for example, from the directory=D:\Logs directory) are parsed by myparser. If the collected logs contain three values that are separated by a semicolon, the parsed events sequentially receive the field_name1, field_name2 and field_name3 names.

To parse the following CSV log:

"United States","USA","North America","High income: OECD","Fiscal year end: September 30; reporting period for national accounts data: CY."

Define the CSV parser configuration:

fields=country_name, country_code, region, income_group, special_notes

The CSV parser returns the following fields:

country_name=United States
region=North America
income_group=High income: OECD
special_notes=Fiscal year end: September 30; reporting period for national accounts data: CY.