Skip to main content
Version: Tenzir v4.9


Reads and writes JSON.



json [--schema=<schema>] [--selector=<field[:prefix]>] [--unnest-separator=<string>]
[--no-infer] [--ndjson] [--raw]


json [-c|--compact-output] [-C|--color-output] [-M|--monochrome-output]
[--omit-nulls] [--omit-empty-objects] [--omit-empty-lists] [--omit-empty]


The json format provides a parser and printer for JSON and line-delimited JSON objects.

--schema=<schema> (Parser)

Provide the name of a schema to be used by the parser. If the schema uses the blob type, then the JSON parser expects base64-encoded strings.

The --schema option is incompatible with the --selector option.

--selector=<field[:prefix]> (Parser)

Designates a field value as schema name with an optional dot-separated prefix.

For example, the Suricata EVE JSON format includes a field event_type that contains the event type. Setting the selector to event_type:suricata causes an event with the value flow for the field event_type to map onto the schema suricata.flow.

The --selector option is incompatible with the --schema option.

--no-infer (Parser)

The JSON parser automatically infers types in the input JSON.

The flag --no-infer toggles this behavior, and requires the user to provide an input schema for the JSON to explicitly parse into, e.g., using the --selector option.

Schema inference happens on a best-effort basis, and is constantly being improved to match Tenzir's type system.

--unnest-separator=<string> (Parser)

A delimiter that, if present in keys, causes values to be treated as values of nested records.

A popular example of this is the Zeek JSON format. It includes the fields id.orig_h, id.orig_p, id.resp_h, and id.resp_p at the top-level. The data is best modeled as an id record with four nested fields orig_h, orig_p, resp_h, and resp_p.

Without an unnest separator, the data looks like this:

"id.orig_h" : "",
"id.orig_p" : 10,
"id.resp_h" : "",
"id.resp_p" : 5

With the unnest separator set to ., Tenzir reads the events like this:

"id" : {
"orig_h" : "",
"orig_p" : 10,
"resp_h" : "",
"resp_p" : 5

--ndjson (Parser)

Treat the input as newline-delimited JSON (NDJSON).

NDJSON requires that exactly one event exists per line. This allows for better error recovery in cases of malformed input, as unlike for the regular JSON parser malformed lines can be skipped.

Popular examples of NDJSON include the Suricata Eve JSON and the Zeek Streaming JSON formats. Tenzir supports suricata and zeek-json parsers out of the box that utilize this mechanism.

--raw (Parser)

Use only the raw JSON types. This means that all strings are parsed as string, irrespective of whether they are a valid ip, duration, etc. Also, since JSON only has one generic number type, all numbers are parsed with the double type.

--arrays-of-objects (Parser)

Parse arrays of objects, with every object in the outermost arrays resulting in one event each. This is particularly useful when interfacing with REST APIs, which often yield large arrays of objects instead of newline-delimited JSON objects.

--c|--compact-output (Printer)

Switch to line-delimited JSON output (JSONL/NDJSON).

--C|--color-output (Printer)

Colorize the output.

The option enables colorizing the output similar to jq by emitting terminal escape sequences that represent colors.

Unlike jq, coloring is currently opt-in. In the future, we will perform TTY detection and colorize the output when write to stdout.

Tenzir honors the NO_COLOR environment variable and won't colorize the output when the variable is present.

--M|--monochrome-output (Printer)

Disables colored output.

This is currently the default. In the future, we will perform TTY detection and colorize the output when write to stdout. Use this option today if you want to avoid an implicit upgrade to colors in the future.

--omit-nulls (Printer)

Strips null fields from the output.


"a": null,
"b": [42, null, 43],
"c": {
"d": null,
"e": 42

With --omit-nulls, this example becomes:

"b": [42, 43],
"c": {
"e": 42

--omit-empty-objects (Printer)

Strips empty objects from the output.


"w": null,
"x": {},
"y": {
"z": {}

With --omit-empty-objects, this example becomes:

"w": 42,

--omit-empty-lists (Printer)

Strips empty lists from the output.


"m": [],
"n": {},
"o": {
"p": []

With --omit-empty-lists, this example becomes:

"m": {},
"o": {}

--omit-empty (Printer)

This options combines all other --omit-* options.


Read JSON from stdin, either NDJSON or tree-structured:

read json

Write JSON without empty fields to a file:

to file /tmp/result.json write json --omit-empty

Print NDJSON to stdout:

write json -c