The parse operator (also called the parse anchor) parses strings according to specified start and stop anchors, and then labels them as fields for use in subsequent aggregation functions in the query such as sorting, grouping, or other functions.
This topic describes how to use the parse anchor UI tool to add parsing to a query and provides details on the structure of the parse anchor operator.
parse anchor UI tool
You can use the parse anchor UI tool to highlight the message text to parse, identify parsing fields, and perform the parsing action.
To parse using the parse anchor tool:
- Run a search.
- In the search results, find a message with the text you want to parse.
- Highlight the text, right-click, and select Parse the selected text.
The Parse Text dialog box opens and displays the text you highlighted.
- Select the text for the first parsing field, and click Click to extract this value.
The text you highlighted is replaced by an asterisk (*).
In this example screenshot, GET is the parsing anchor, and the highlighted text that follows is the first parsing field.
- Enter a name (no spaces) for the parsing field in the Fields area.
- If you want to parse additional fields, add a comma after the field name, and repeat the parsing action. The following screenshot shows three parsed fields: url, status_code, and size (in that order). Notice that the three fields correspond to the three asterisks in the parse text.
- Click Submit.
The Search page reopens to show the parse operator you just constructed added to the search.
- Click Start to display the search results, which now show the parsed message.
This section provides details on the parse anchor operator. You can create parse anchor queries using the UI tool described above, or construct your own queries using the information in this section.
... | parse "start_anchor*stop_anchor" as fieldname | ...
... | parse "start_anchor*stop_anchor" as fieldname no drop| ...
... | parse field=<field_name> "start_anchor*stop_anchor" as fieldname | ...
parse field=fieldnameoption allows you to specify a field to parse other than the default message. For details, see Parse field.
parse nodropoption forces results to also include messages that do not match any segment of the parse term. For details, see Parse nodrop.
- User-created fields, such as extracted or parsed fields, can be named using alphanumeric characters and underscores (
_). Fields must start with an alphanumeric character.
- If no field is specified, the entire text of incoming messages is used.
- A wildcard is used as a placeholder for the extracted field. Wildcards must be separated by a space or other character.
**is not valid.
- The number of wildcards in the pattern string must match the number of variables.
- Multiple expressions are allowed for a single parse operator.
- Can be used with parse regex operator.
Sample log message:
Aug 2 04:06:08: host=10.1.1.124: local/ssl2 notice mcpd: Userfirstname.lastname@example.org: severity=warning: 01070638:5: Pool member 172.31.51.22:0 monitor status down.
In the following examples, the start_anchor is "user=" and the stop_anchor is ":", which ends the email address. The asterisk (*) is the glob representing the parsed term. The examples create a new field for each message named "user" and that field will contain the value of the email address, in this case email@example.com.
... | parse "user=*:" as user
The parse operator also allows you to extract multiple fields in one command:
... | parse "user=*: severity=*:" as user, severity | ...
This example creates two fields from the sample log message:
Name Fields with Special Characters
You can create field names that contain special characters, for example, spaces, dashes, and backslashes or forward slashes, using the following syntax:
... | parse "<string>" as %"<field name with special characters>"
For example, this query will allow you to parse the phrase "Class ID", including the space:
... | parse "[Classification:*]" as %"Class ID"
Use Line Breaks as an Anchor
If your logs are delivered in a multi-line format, you may want to parse up until a line break in the message. In order to do so, use the following regular expressions as a stop anchor on the line break:
Linux Logs- "\n"
Windows Logs- "\r"
For example, if we have the following message in our logs:
12:08:10,651 INFO sample_server ReportEmailer:178 - DEBUG SENDING MESSAGE: To: firstname.lastname@example.org Subject: New line Breaks in Message
to get the "To:" address, you can use the following queries to get the address:
... | parse "To:*\n" as ToAddress
... | parse "To:*\r" as ToAddress
which returns email@example.com in the ToAddress column.