Command line and exit codes

The tdbio tool (package TdbIO.Cli) reports the structure of a file. It prints text or JSON to standard output, writes messages and warnings to standard error and tells success from failure by its exit code. It uses only the public API of the library.

Usage: tdbio info <file> [--json] [--encoding ascii|utf-8|cp1251|cp1252]
       tdbio tables <file> [--json]
       tdbio text <file> [--json]
       tdbio --help | -h | --version

Commands

Command Works on Output
tdbio info <file> every supported kind key: value lines: the kind, its main values and the number of warnings.
tdbio tables <file> model databases (.db1), DBV@ files A tab-separated header line, then one line per table section.
tdbio text <file> model databases (.db1) A tab-separated header line, then one line per text chain, then one line per orphan row.
tdbio --help, tdbio -h The usage text, on standard output.
tdbio --version tdbio <version>, for example tdbio 0.1.0.

The command comes first; options may come before or after the file. The kind of the file is found from its content, not from its extension (reports, which have no signature, are recognised by the .xsr file name).

Options

Option Commands Effect
--json all Write one indented JSON object (UTF-8) instead of text.
--encoding <name> info Read a template or report with this encoding instead of detecting it: ascii, utf-8, cp1251 or cp1252. The encoding is strict: a byte it does not define is an input error. Given for any other kind of file, it is a usage error.

An option given twice is a usage error.

Exit codes

Code Name Meaning
0 done The command finished and its output was written. Warnings do not change the code.
1 usage The command line is wrong: no command, an unknown command or option, no file, an empty file name, an extra argument, an option repeated, --encoding without a value or with an unknown one, or --encoding for a file that is not a template or report. Two lines on standard error: tdbio: <problem> and Run 'tdbio --help' for usage.
2 input The file cannot be used: it does not exist or cannot be read, is not recognised, is cut or malformed, is over a limit, is not supported (a drawing file, UTF-16 text), or is of a kind the command does not handle (for example tables on a .nc1 file). One line on standard error, tdbio: <problem>, no stack trace.

The whole output is built before anything is written, so a failing command writes nothing to standard output. A message about malformed data ends with the offset of the problem, for example tdbio: not a recognised Tekla file at offset 0x0.

Warnings are written after the output, one line each: tdbio: warning: <warning>. They name an offset or a line, never text from the file.

info

The first key is kind, the name of the TdbFileKind value. In text output the last key is warnings, the number of warnings; in JSON output warnings is the list of warning lines. A value that is not there is - in text and null in JSON.

Kind Keys
Db1 version, modelGuid, idMapEntries, idMapField65 (hex in text), idMapField73, textSections, textChains, saveStamps, lastSavedAt (the latest save stamp date)
Db2 version, modelGuid, records, rows (the sum of the row counts)
GuidMapper pairs
IdRanges next, ranges
History pageSize, pageCount, length
Dbv name, records (GUID records), sections, rows (the sum over all sections)
Nc1 blocks (the block codes in order), mark, material, quantity, profile, length
Template encoding, rootKeyword, instance, type, version (the values of the root properties of these names), blocks (the number of blocks below the root), maxDepth (the deepest block level; the root is level 0)
Report encoding, lines, pages (the number of form feeds), endsWithLineBreak

For a model database info also reads the text sections, so a malformed text section makes it fail with exit code 2.

Example, on a synthetic DSTV NC file:

$ tdbio info part.nc1
kind: Nc1
blocks: ST,AK,EN
mark: A1
material: GRADE-X
quantity: 3
profile: PL12
length: 100
warnings: 0
$ tdbio info part.nc1 --json
{
  "kind": "Nc1",
  "blocks": [
    "ST",
    "AK",
    "EN"
  ],
  "mark": "A1",
  "material": "GRADE-X",
  "quantity": 3,
  "profile": "PL12",
  "length": 100,
  "warnings": []
}

With --encoding, encoding shows the web name of the encoding used, for example windows-1252 for cp1252.

tables

Columns, separated by tabs:

Column Value
ordinal The position of the section in the file, from 0.
offset Where the section starts (the key when there is one, otherwise the separator), hex.
key The raw key stored before the section, or -. Always - for DBV@ files.
rowSize The number of body bytes in every row.
descriptors The raw descriptor words, separated by commas.
rows The number of rows.
end The offset just after the section, hex.

Offsets refer to the unpacked file. Row bodies are never printed.

$ tdbio tables model.db1
ordinal	offset	key	rowSize	descriptors	rows	end
0	0x7D	-	8	1,0	2	0xB4
1	0xB4	42	4	1	1	0xDA

With --json the object has kind, sections (one object per section with the column names as properties; offsets are decimal numbers, key is null when absent and descriptors is an array) and warnings.

text

Columns, separated by tabs:

Column Value
section The index of the text section, from 0.
headId The id of the head row of the chain.
rows The number of rows in the chain.
length The number of text bytes.
stamp true for a save stamp, else false.
savedAt The date and time of a save stamp, yyyy-MM-ddTHH:mm:ss, or -.
text The text bytes, escaped: printable ASCII as is, \\, \t, \n, and \xNN for every other byte. - for a save stamp.

The text of a save stamp is never printed: besides the date it can name people or machines. Rows that belong to no chain follow the chains, one line each: the section index, -, orphan and the row id.

$ tdbio text model.db1
section	headId	rows	length	stamp	savedAt	text
0	1	1	6	false	-	Grid A
0	10	2	42	true	2026-02-01T10:20:30	-

With --json the object has kind, chains (one object per chain with the column names as properties; savedAt and text are null when absent), orphans (objects with section and id) and warnings.

JSON output

Numbers are decimal, GUIDs use the D format (xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx) and dates the format yyyy-MM-ddTHH:mm:ss. The input path is never written. Warnings are both in the warnings array and on standard error.

Text output

Each value fits on one line: control characters in text taken from a file are written as \xNN. Lists are joined with commas. Numbers use the invariant culture.

All examples on this page come from small synthetic files; the values are made up.