tensogram ls

Lists messages in a Tensogram file, showing metadata in tabular or JSON format.

Usage

tensogram ls [OPTIONS] [FILES]...

Options

Examples

# List all messages with default columns
tensogram ls forecast.tgm

# Only temperature fields
tensogram ls forecast.tgm -w "mars.param=2t"

# Temperature or wind
tensogram ls forecast.tgm -w "mars.param=2t/10u/10v"

# Exclude ensemble members
tensogram ls forecast.tgm -w "mars.type!=em"

# Show only date and step columns
tensogram ls forecast.tgm -p "mars.date,mars.step"

# JSON output (one object per line, good for jq)
tensogram ls forecast.tgm -j | jq '.["mars.param"]'

Where Clause Syntax

The -w flag accepts a single expression:

key=value           # exact match
key=v1/v2/v3        # OR — matches any of v1, v2, v3
key!=value          # not equal
key!=v1/v2          # not any of v1, v2

Key format: namespace.field for namespaced keys (e.g. mars.param) or just field for top-level keys (e.g. version).

Missing key: For key=value, a missing key is treated as non-matching. For key!=value, a missing key passes the filter.

Only one -w expression can be specified per command. To apply multiple filters, pipe commands:

tensogram ls forecast.tgm -w "mars.type=fc" | grep "2t"

Pick Keys

The -p flag selects which metadata columns to display. Keys use the same dot-notation as -w:

tensogram ls forecast.tgm -p "mars.date,mars.step,mars.param"

Without -p, all available metadata keys are shown.

Default Table Output

mars.date   mars.step  mars.param  mars.type  shape
20260401    0          2t          fc         [721, 1440]
20260401    0          10u         fc         [721, 1440]
20260401    0          10v         fc         [721, 1440]
20260401    6          2t          fc         [721, 1440]
...

JSON Output

With -j, each matching message is printed as a JSON object on its own line:

{"mars.date": "20260401", "mars.step": "0", "mars.param": "2t", "shape": "[721, 1440]"}
{"mars.date": "20260401", "mars.step": "0", "mars.param": "10u", "shape": "[721, 1440]"}

This is compatible with jq, grep, and any tool that processes newline-delimited JSON.