miller/doc/content-for-file-formats.html

POKI_PUT_TOC_HERE

<p/>
<button style="font-weight:bold;color:maroon;border:0" onclick="bodyToggler.expandAll();" href="javascript:;">Expand all sections</button>
<button style="font-weight:bold;color:maroon;border:0" onclick="bodyToggler.collapseAll();" href="javascript:;">Collapse all sections</button>

<h1>Overview</h1>
<button style="font-weight:bold;color:maroon;border:0" padding=0 onclick="bodyToggler.toggle('body_section_toggle_overview');" href="javascript:;">Toggle section visibility</button>
<div id="body_section_toggle_overview" style="display: block">

<p/>Miller handles name-indexed data using several formats: some you probably
know by name, such as CSV, TSV, and JSON &mdash; and other formats you&rsquo;re
likely already seeing and using in your structured data. Additionally, Miller gives
you the option of including comments within your data.

</div>
<h1>Examples</h1>
<button style="font-weight:bold;color:maroon;border:0" padding=0 onclick="bodyToggler.toggle('body_section_toggle_examples');" href="javascript:;">Toggle section visibility</button>
<div id="body_section_toggle_examples" style="display: block">

POKI_RUN_COMMAND{{mlr --usage-data-format-examples}}HERE

</div>
<h1>CSV/TSV/ASV/USV/etc.</h1>
<button style="font-weight:bold;color:maroon;border:0" padding=0 onclick="bodyToggler.toggle('body_section_toggle_csv_tsv_etc');" href="javascript:;">Toggle section visibility</button>
<div id="body_section_toggle_csv_tsv_etc" style="display: block">

<p/>
When <code>mlr</code> is invoked with the <code>--csv</code> or <code>--csvlite</code> option,
key names are found on the first record and values are taken from subsequent
records.  This includes the case of CSV-formatted files.  See
POKI_PUT_LINK_FOR_PAGE(record-heterogeneity.html)HERE for how Miller handles
changes of field names within a single data stream.

<p/> Miller has record separator <code>RS</code> and field separator <code>FS</code>,
just as <code>awk</code> does.  For TSV, use <code>--fs tab</code>; to convert TSV to
CSV, use <code>--ifs tab --ofs comma</code>, etc.  (See also
POKI_PUT_LINK_FOR_PAGE(reference.html#Record/field/pair_separators)HERE.)

<p/> <b>TSV (tab-separated values):</b> the following are synonymous pairs:
<ul>
<li/> <code>--tsv</code> and <code>--csv --fs tab</code>
<li/> <code>--itsv</code> and <code>--icsv --ifs tab</code>
<li/> <code>--otsv</code> and <code>--ocsv --ofs tab</code>
<li/> <code>--tsvlite</code> and <code>--csvlite --fs tab</code>
<li/> <code>--itsvlite</code> and <code>--icsvlite --ifs tab</code>
<li/> <code>--otsvlite</code> and <code>--ocsvlite --ofs tab</code>
</ul>

<p/> <b>ASV (ASCII-separated values):</b> the flags
<code>--asv</code>,
<code>--iasv</code>,
<code>--oasv</code>,
<code>--asvlite</code>,
<code>--iasvlite</code>, and
<code>--oasvlite</code> are analogous except they use ASCII FS and RS 0x1f and 0x1e, respectively.

<p/> <b>USV (Unicode-separated values):</b> likewise, the flags
<code>--usv</code>,
<code>--iusv</code>,
<code>--ousv</code>,
<code>--usvlite</code>,
<code>--iusvlite</code>, and
<code>--ousvlite</code> use Unicode FS and RS U+241F (UTF-8 0x0xe2909f) and U+241E (UTF-8 0xe2909e), respectively.

<p/>Miller&rsquo;s <code>--csv</code> flag supports RFC-4180 CSV (<a href="https://tools.ietf.org/html/rfc4180">
https://tools.ietf.org/html/rfc4180</a>). This includes CRLF line-terminators by default, regardless
of platform.

<p/>Here are the differences between CSV and CSV-lite:

<ul>

<li/>CSV supports <a href="https://tools.ietf.org/html/rfc4180">
RFC-4180</a>)-style double-quoting, including the ability to have commas and/or
LF/CRLF line-endings contained within an input field; CSV-lite does not.

<li/>CSV does not allow heterogeneous data; CSV-lite does (see also <a
href="record-heterogeneity.html">here</a>).

<li/>The CSV-lite input-reading code is fractionally more efficient than the CSV
input-reader.

</ul>

<p/>Here are things they have in common:

<ul>

<li/>The ability to specify record/field separators other than the default,
e.g. CR-LF vs. LF, or tab instead of comma for TSV, and so on.

<li/>The <code>--implicit-csv-header</code> flag for input and the
<code>--headerless-csv-output</code> flag for output.

</ul>

</div>
<h1>DKVP: Key-value pairs</h1>
<button style="font-weight:bold;color:maroon;border:0" padding=0 onclick="bodyToggler.toggle('body_section_toggle_dkvp');" href="javascript:;">Toggle section visibility</button>
<div id="body_section_toggle_dkvp" style="display: block">

<p/>
Miller&rsquo;s default file format is DKVP, for <b>delimited key-value pairs</b>. Example:

POKI_RUN_COMMAND{{mlr cat data/small}}HERE

Such data are easy to generate, e.g. in Ruby with
POKI_CARDIFY(puts "host=#{hostname},seconds=#{t2-t1},message=#{msg}")HERE
POKI_CARDIFY{{puts mymap.collect{|k,v| "#{k}=#{v}"}.join(',')}}HERE
or <code>print</code> statements in various languages, e.g.
POKI_CARDIFY(echo "type=3,user=$USER,date=$date\n";)HERE
POKI_CARDIFY{{logger.log("type=3,user=$USER,date=$date\n");}}HERE

<p/>Fields lacking an IPS will have positional index (starting at 1) used as
the key, as in NIDX format. For example, <code>dish=7,egg=8,flint</code> is parsed
as <code>"dish" =&gt; "7", "egg" =&gt; "8", "3" =&gt; "flint"</code> and
<code>dish,egg,flint</code> is parsed as <code>"1" =&gt; "dish", "2" =&gt; "egg", "3"
=&gt; "flint"</code>.

<p/> As discussed in POKI_PUT_LINK_FOR_PAGE(record-heterogeneity.html)HERE,
Miller handles changes of field names within the same data stream. But using
DKVP format this is particularly natural. One of my favorite use-cases for
Miller is in application/server logs, where I log all sorts of lines such as

<p/>
<div class="pokipanel">
<pre>
resource=/path/to/file,loadsec=0.45,ok=true
record_count=100, resource=/path/to/file
resource=/some/other/path,loadsec=0.97,ok=false
</pre>
</div>
<p/>

etc. and I just log them as needed. Then later, I can use <code>grep</code>, <code>mlr --opprint group-like</code>, etc.
to analyze my logs.

<p/>See POKI_PUT_LINK_FOR_PAGE(reference.html)HERE regarding how to specify separators other than
the default equals-sign and comma.

</div>
<h1>NIDX: Index-numbered (toolkit style)</h1>
<button style="font-weight:bold;color:maroon;border:0" padding=0 onclick="bodyToggler.toggle('body_section_toggle_nidx');" href="javascript:;">Toggle section visibility</button>
<div id="body_section_toggle_nidx" style="display: block">

With <code>--inidx --ifs ' ' --repifs</code>, Miller splits lines on whitespace and
assigns integer field names starting with 1. This recapitulates Unix-toolkit
behavior.

<p/> Example with index-numbered output:

<table><tr> <td>
POKI_RUN_COMMAND{{cat data/small}}HERE
</td> <td>
POKI_RUN_COMMAND{{mlr --onidx --ofs ' ' cat data/small}}HERE
</td> </tr></table>

<p/> Example with index-numbered input:

<table><tr> <td>
POKI_RUN_COMMAND{{cat data/mydata.txt}}HERE
</td> <td>
POKI_RUN_COMMAND{{mlr --inidx --ifs ' ' --odkvp cat data/mydata.txt}}HERE
</td> </tr></table>

<p/> Example with index-numbered input and output:

<table><tr> <td>
POKI_RUN_COMMAND{{cat data/mydata.txt}}HERE
</td> <td>
POKI_RUN_COMMAND{{mlr --nidx --fs ' ' --repifs cut -f 2,3 data/mydata.txt}}HERE
</td> </tr></table>

</div>
<h1>Tabular JSON</h1>
<button style="font-weight:bold;color:maroon;border:0" padding=0 onclick="bodyToggler.toggle('body_section_toggle_json');" href="javascript:;">Toggle section visibility</button>
<div id="body_section_toggle_json" style="display: block">

<p/>JSON is a format which supports arbitrarily deep nesting of
&ldquo;objects&rdquo; (hashmaps) and &ldquo;arrays&rdquo; (lists), while Miller
is a tool for handling <span class="boldmaroon">tabular data</span> only. This means
Miller cannot (and should not) handle arbitrary JSON. (Check out <a
href="http://stedolan.github.io/jq/">jq</a>.)

<p/>But if you have tabular data represented in JSON then Miller can handle that for you.

<h2>Single-level JSON objects</h2>

An <span class="boldmaroon">array of single-level objects</span> is, quite simply,
<span class="boldmaroon">a table:</span>

POKI_RUN_COMMAND{{mlr --json head -n 2 then cut -f color,shape data/json-example-1.json}}HERE
POKI_RUN_COMMAND{{mlr --json --jvstack head -n 2 then cut -f color,u,v data/json-example-1.json}}HERE
POKI_RUN_COMMAND{{mlr --ijson --opprint stats1 -a mean,stddev,count -f u -g shape data/json-example-1.json}}HERE

<h2>Nested JSON objects</h2>

<p/>Additionally, Miller can <span class="boldmaroon">tabularize nested objects by concatentating keys:</span>

POKI_RUN_COMMAND{{mlr --json --jvstack head -n 2 data/json-example-2.json}}HERE
POKI_RUN_COMMAND{{mlr --ijson --opprint head -n 4 data/json-example-2.json}}HERE

<p/>Note in particular that as far as Miller&rsquo;s <code>put</code> and <code>filter</code>, as well as other
I/O formats, are concerned, these are simply field names with colons in them:

POKI_RUN_COMMAND{{mlr --json --jvstack head -n 1 then put '${values:uv} = ${values:u} * ${values:v}' data/json-example-2.json}}HERE

<h2>Arrays</h2>

<p/>Arrays aren&rsquo;t supported in Miller&rsquo;s
<code>put</code>/<code>filter</code> DSL. By default, JSON arrays are read in as
integer-keyed maps.

<p/> Suppose you have arrays like this in our input data:

POKI_RUN_COMMAND{{cat data/json-example-3.json}}HERE

<p/> Then integer indices (starting from 0 and counting up) are used as map keys:

POKI_RUN_COMMAND{{mlr --ijson --oxtab cat data/json-example-3.json}}HERE

<p/> When the data are written back out as JSON, field names are re-expanded as above,
but what were arrays on input are now maps on output:

POKI_RUN_COMMAND{{mlr --json --jvstack cat data/json-example-3.json}}HERE

<p/> This is non-ideal, but it allows Miller (5.x release being latest as of
this writing) to handle JSON arrays at all.

<p/>You might also use <code>mlr --json-skip-arrays-on-input</code> or <code>mlr
--json-fatal-arrays-on-input</code>.

To truly handle JSON, please use a JSON-processing tool such as <a
href="http://stedolan.github.io/jq/">jq</a>.

<h2>Formatting JSON options</h2>

<p/>JSON isn&rsquo;t a parameterized format, so <code>RS</code>, <code>FS</code>,
<code>PS</code> aren&rsquo;t specifiable. Nonetheless, you can do the following:

<ul>

<li/>Use <code>--jvstack</code> to pretty-print JSON objects with multi-line
(vertically stacked) spacing. By default, each Miller record (JSON object) is
one per line.

<li/>Keystroke-savers:
<code>--jsonx</code> simply means <code>--json --jvstack</code>, and
<code>--ojsonx</code> simply means <code>--ojson --jvstack</code>.

<li/>Use <code>--jlistwrap</code> to print the sequence of JSON objects wrapped in
an outermost <code>[</code> and <code>]</code>. By default, these aren&rsquo;t printed.

<li/>Use <code>--jquoteall</code> to double-quote all object values. By default,
integers, floating-point numbers, and booleans <code>true</code> and <code>false</code>
are not double-quoted when they appear as JSON-object keys.

<li/>Use <code>--jflatsep yourstringhere</code> to specify the string used for
key concatenation: this defaults to a single colon.

<li/>Use <code>--jofmt</code> to force Miller to apply the global <code>--ofmt</code>
to floating-point values.  First note: please use sprintf-style codes for
double precision, e.g. ending in <code>%lf</code>, <code>%le</code>, or <code>%lg</code>.
Miller floats are double-precision so behavior using <code>%f</code>, <code>%d</code>,
etc. is undefined.  Second note: <code>0.123</code> is valid JSON; <code>.123</code> is
not. Thus this feature allows you to emit JSON which may be unparseable by
other tools.

</ul>

<p/>Again, please see <a href="http://stedolan.github.io/jq/">jq</a> for a
truly powerful, JSON-specific tool.

<h2>JSON non-streaming</h2>

<p/>The JSON parser Miller uses does not return until all input is parsed: in
particular this means that, unlike for other file formats, Miller does not (at
present) handle JSON files in <code>tail -f</code> contexts.

</div>
<h1>PPRINT: Pretty-printed tabular</h1>
<button style="font-weight:bold;color:maroon;border:0" padding=0 onclick="bodyToggler.toggle('body_section_toggle_pprint');" href="javascript:;">Toggle section visibility</button>
<div id="body_section_toggle_pprint" style="display: block">

<p/>
Miller&rsquo;s pretty-print format is like CSV, but column-aligned.  For example, compare

<table><tr><td>
POKI_RUN_COMMAND{{mlr --ocsv cat data/small}}HERE
</td>
<td>
POKI_RUN_COMMAND{{mlr --opprint cat data/small}}HERE
</td></tr></table>

Note that while Miller is a line-at-a-time processor and retains input lines in
memory only where necessary (e.g. for sort), pretty-print output requires it to
accumulate all input lines (so that it can compute maximum column widths)
before producing any output. This has two consequences: (a) pretty-print output
won&rsquo;t work on <code>tail -f</code> contexts, where Miller will be waiting for
an end-of-file marker which never arrives; (b) pretty-print output for large
files is constrained by available machine memory.

<p/> See POKI_PUT_LINK_FOR_PAGE(record-heterogeneity.html)HERE for how Miller
handles changes of field names within a single data stream.

<p/> For output only (this isn&rsquo;t supported in the input-scanner as of 5.0.0)
you can use <code>--barred</code> with pprint output format:

POKI_RUN_COMMAND{{mlr --opprint --barred cat data/small}}HERE

</div>
<h1>XTAB: Vertical tabular</h1>
<button style="font-weight:bold;color:maroon;border:0" padding=0 onclick="bodyToggler.toggle('body_section_toggle_xtab');" href="javascript:;">Toggle section visibility</button>
<div id="body_section_toggle_xtab" style="display: block">

<p/>This is perhaps most useful for looking a very wide and/or multi-column
data which causes line-wraps on the screen (but see also <a
href="https://github.com/twosigma/ngrid">https://github.com/twosigma/ngrid</a>
for an entirely different, very powerful option). Namely:

<table><tr> <td>
POKI_INCLUDE_ESCAPED(data/system-file-opprint-example.txt)HERE
</td></tr> <tr><td>
POKI_INCLUDE_ESCAPED(data/system-file-oxtab-example.txt)HERE
</td></tr> <tr><td>
POKI_INCLUDE_ESCAPED(data/system-file-ojson-example.txt)HERE
</td> </tr></table>

</div>
<h1>Markdown tabular</h1>
<button style="font-weight:bold;color:maroon;border:0" padding=0 onclick="bodyToggler.toggle('body_section_toggle_markdown_tabular');" href="javascript:;">Toggle section visibility</button>
<div id="body_section_toggle_markdown_tabular" style="display: block">

<p/>Markdown format looks like this:

POKI_RUN_COMMAND{{mlr --omd cat data/small}}HERE

which renders like this when dropped into various web tools (e.g. github comments):

<p/>
<img src="pix/omd.png"/>

<p/> As of Miller 4.3.0, markdown format is supported only for output, not input.

</div>
<h1>Data-conversion keystroke-savers</h1>
<button style="font-weight:bold;color:maroon;border:0" padding=0 onclick="bodyToggler.toggle('body_section_toggle_x2y');" href="javascript:;">Toggle section visibility</button>
<div id="body_section_toggle_x2y" style="display: block">

<p/>While you can do format conversion using <code>mlr --icsv --ojson cat myfile.csv</code>,
there are also keystroke-savers for this purpose, such as <code>mlr --c2j cat myfile.csv</code>.
For a complete list:

POKI_RUN_COMMAND{{mlr --usage-format-conversion-keystroke-saver-options}}HERE

</div>

</div>
<h1>Autodetect of line endings</h1>
<button style="font-weight:bold;color:maroon;border:0" padding=0 onclick="bodyToggler.toggle('body_section_toggle_autodetect_line_endings');" href="javascript:;">Toggle section visibility</button>
<div id="body_section_toggle_autodetect_line_endings" style="display: block">

<p/> Default line endings (<code>--irs</code> and <code>--ors</code>) are <code>'auto'</code>
which means <span class="boldmaroon">autodetect from the input file format</span>, as
long as the input file(s) have lines ending in either LF (also known as
linefeed, <code>'\n'</code>, <code>0x0a</code>, Unix-style) or CRLF (also known as
carriage-return/linefeed pairs, <code>'\r\n'</code>, <code>0x0d 0x0a</code>, Windows
style).

<p/> <span class="boldmaroon">If both IRS and ORS are auto (which is the default) then LF input will
lead to LF output and CRLF input will lead to CRLF output, regardless of the
platform you&rsquo;re running on</span>.

<p/> The line-ending autodetector triggers on the first line ending detected in
the input stream. E.g. if you specify a CRLF-terminated file on the command
line followed by an LF-terminated file then autodetected line endings will be
CRLF.

<p/> If you use <code>--ors {something else}</code> with (default or explicitly
specified) <code>--irs auto</code> then line endings are autodetected on input and
set to what you specify on output.

<p/> If you use <code>--irs {something else}</code> with (default or explicitly
specified) <code>--ors auto</code> then the output line endings used are LF on
Unix/Linux/BSD/MacOSX, and CRLF on Windows.

<p/> See also
POKI_PUT_LINK_FOR_PAGE(reference.html#Record/field/pair_separators)HERE for
more information about record/field/pair separators.

</div>
<h1>Comments in data</h1>
<button style="font-weight:bold;color:maroon;border:0" padding=0 onclick="bodyToggler.toggle('body_section_toggle_comments_in_data');" href="javascript:;">Toggle section visibility</button>
<div id="body_section_toggle_comments_in_data" style="display: block">

<p/>You can include comments within your data files, and either have them ignored, or passed directly
through to the standard output as soon as they are encountered:

POKI_RUN_COMMAND{{mlr --usage-comments-in-data}}HERE

<p/>Examples:

POKI_RUN_COMMAND{{cat data/budget.csv}}HERE
POKI_RUN_COMMAND{{mlr --skip-comments --icsv --opprint sort -nr quantity data/budget.csv}}HERE
POKI_RUN_COMMAND{{mlr --pass-comments --icsv --opprint sort -nr quantity data/budget.csv}}HERE

</div>