DOKK / manpages / debian 10 / vnlog / vnl-filter.1.en
VNL-FILTER(1) vnlog VNL-FILTER(1)

vnl-filter - filters vnlogs to select particular rows, fields

 $ cat run.vnl
 # time x   y   z   temperature
 3      1   2.3 4.8 30
 4      1.1 2.2 4.7 31
 6      1   2.0 4.0 35
 7      1   1.6 3.1 42
 $ <run.vnl vnl-filter -p x,y,z | vnl-align
 # x  y   z
 1   2.3 4.8
 1.1 2.2 4.7
 1   2.0 4.0
 1   1.6 3.1
 $ <run.vnl vnl-filter -p i=NR,time,'dist=sqrt(x*x + y*y + z*z)' | vnl-align
 # i time   dist
 1   3    5.41572
 2   4    5.30471
 3   6    4.58258
 4   7    3.62905
 $ <run.vnl vnl-filter 'temperature >= 35' | vnl-align
 # time x  y   z  temperature
 6      1 2.0 4.0 35
 7      1 1.6 3.1 42
 $ <run.vnl vnl-filter --eval '{s += temperature} END { print "mean temp: " s/NR}'
 mean temp: 34.5
 $ <run.vnl vnl-filter -p x,y | feedgnuplot --terminal 'dumb 80,30' --unset grid --domain --lines --exit
   2.3 +---------------------------------------------------------------------+
       |           +          +          ***************         +           |
       |                                                **************       |
       |                                                              *******|
   2.2 |-+                                                       ************|
       |                                                 ********            |
       |                                         ********                    |
   2.1 |-+                              *********                          +-|
       |                        ********                                     |
       |                ********                                             |
       |            ****                                                     |
     2 |-+         *                                                       +-|
       |           *                                                         |
       |           *                                                         |
       |           *                                                         |
   1.9 |-+         *                                                       +-|
       |           *                                                         |
       |           *                                                         |
       |           *                                                         |
   1.8 |-+         *                                                       +-|
       |           *                                                         |
       |           *                                                         |
   1.7 |-+         *                                                       +-|
       |           *                                                         |
       |           *                                                         |
       |           *          +           +           +          +           |
   1.6 +---------------------------------------------------------------------+
      0.98         1         1.02        1.04        1.06       1.08        1.1

This tool is largely a frontend for awk to operate on vnlog files. Vnlog is both an input and an output. This tool makes it very simple to select specific rows and columns for output and to manipulate the data in various ways.

This is a UNIX-style tool, so the input/output of this tool is strictly STDIN/STDOUT. Furthermore, in its usual form this tool is a filter, so the format of the output is exactly the same as the format of the input. The exception to this is when using "--eval", in which the output is dependent on whatever expression we're evaluating.

This tool is convenient to process both stored data or live data; in the latter case, it's very useful to pipe the streaming output to "feedgnuplot --stream" to get a realtime visualization of the incoming data.

This tool reads enough of the input file to get a legend, at which point it constructs an awk program to do the main work, and execs to awk (it's possible to use perl as well, but this isn't as fast).

The input/output data is vnlog: a plain-text table of values. Any lines beginning with "#" are treated as comments, and are passed through. The first line that begins with "#" but not "##" or "#!" is a legend line. After the "#", follow whitespace-separated field names. Each subsequent line is whitespace-separated values matching this legend. For instance, this is a valid vnlog file:

 #!/usr/bin/something
 ## more comments
 # x y z
 -0.016107 0.004362 0.005369
 -0.017449 0.006711 0.006711
 -0.018456 0.014093 0.006711
 -0.017449 0.018791 0.006376

"vnl-filter" uses this format for both the input and the output. The comments are preserved, but the legend is updated to reflect the fields in the output file.

A string "-" is used to indicate an undefined value, so this is also a valid vnlog file:

 # x y z
 1 2 3
 4 - 6
 - - 7

To select specific columns, pass their names to the "-p" option (short for "--print" or "--pick", which are synonyms). In its simplest form, to grab only columns "x" and "y", do

 vnl-filter -p x,y

See the detailed description of "-p" below for more detail.

To select specific rows, we use matches expressions. Anything on the "vnl-filter" commandline and not attached to any "--xxx" option is such an expression. For instance

 vnl-filter 'size > 10'

would select only those rows whose "size" column contains a value > 10. See the detailed description of matches expressions below for more detail.

By default, the parsing of arguments and the legend happens in perl, which then constructs a simple awk script, and invokes "mawk" to actually read the data and to process it. This is done because awk is lighter weight and runs faster, which is important because our data sets could be quite large. We default to "mawk" specifically, since this is a simpler implementation than "gawk", and runs much faster. If for whatever reason we want to do everything with perl, this can be requested with the "--perl" option.

For convenience we support several special functions in any expression passed on to awk or perl (named expressions, matches expressions, "--eval" strings). These are

  • rel(x) returns value of "x" relative to the first value of "x". For instance we might want to see the time or position relative to the start, not relative to some absolute beginning. Example:

     $ cat tst.vnl
     # time x
     100    200
     101    212
     102    209
     $ <tst.vnl vnl-filter -p 't=rel(time),x=rel(x)
     # t x
     0 0
     1 12
     2 9
        
  • diff(x) returns the difference between the current value of "x" and the previous value of "x". The first row will always be 0. Example:

     $ <tst.vnl vnl-filter -p x,'d1=diff(x),d2=diff(diff(x))' | vnl-align
     # x d1 d2
       1  0  0
       8  7  7
      27 19 12
      64 37 18
     125 61 24
        
  • sum(x) returns the cumulative sum of "x". As diff(x) can be thought of as a derivative, sum(x) can be thought of as an integral. So "diff(sum(x))" would return the same value as "x" (except for the first row; "diff()" always returns 0 for the first row).

    Example:

     $ <tst.vnl vnl-filter -p 'x,s=sum(x),ds=diff(sum(x))' | vnl-align
     # x  s   ds
       1   1   0
       8   9   8
      27  36  27
      64 100  64
     125 225 125
        
  • prev(x) returns the previous value of "x". One could construct "sum()" and "rel()" using this, if they weren't already available.

-p|--print|--pick expr

These option provide the mechanism to select specific columns for output. For instance to pull out columns called "lat", "lon", and any column whose name contains the string "feature_", do

 vnl-filter -p lat,lon,'feature_.*'

or, equivalently

 vnl-filter --print lat --print lon --print 'feature_.*'

We look for exact column name matches first, and if none are found, we try a regex. If there was no column called exactly "feature_", then the above would be equivalent to

 vnl-filter -p lat,lon,feature_

This mechanism is much more powerful than just selecting columns. First off, we can rename chosen fields:

 vnl-filter -p w=feature_width

would pick the "feature_width" field, but the resulting column in the output would be named "w". When renaming a column in this way regexen are not supported, and exact field names must be given. But the string to the right of the "=" is passed on directly to awk (after replacing field names with column indices), so any awk expression can be used here. For instance to compute the length of a vector in separate columns "x", "y", and "z" you can do:

 vnl-filter -p 'l=sqrt(x*x + y*y + z*z)'

A single column called "l" would be produced.

We can also exclude columns by preceding their name with "!". This works like you expect. Rules:

  • The pick/exclude directives are processed in order given to produce the output picked-column list
  • If the first "-p" item is an exclusion, we implicitly pick all the columns prior to processing the "-p".
  • The exclusion expressions match the output column names, not the input names.
  • We match the exact column names first. If that fails, we match as a regex

Example. To grab all the columns except the temperature(s) do this:

 vnl-filter -p !temperature

To grab all the columns that describe something about a robot (columns whose names have the string "robot_" in them), but not its temperature (i.e. not "robot_temperature"), do this:

 vnl-filter -p robot_,!temperature

--has a,b,c,...

Used to select records (rows) that have a non-empty value in a particular field (column). A null value in a column is designated with a single "-". If we want to select only records that have a value in the "x" column, we pass "--has x". To select records that have data for all of a given set of columns, the "--has" option can be repeated, or these multiple columns can be given in a whitespace-less comma-separated list. For instance if we want only records that have data in both columns "x" and "y" we can pass in "--has x,y" or "--has x --has y". If we want to combine multiple columns in an or (select rows that have data in any of a given set of columns), use a matches expression, as documented below.

If we want to select a column and pick only rows that have a value in this column, a shorthand syntax exists:

 vnl-filter --has col -p col

is equivalent to

 vnl-filter -p +col

Anything on the commandline not attached to any "--xxx" option is a matches expression. These are used to select particular records (rows) in a data file. For each row, we evaluate all the expressions. If all the expressions evaluate to true, that row is output. This expression is passed directly to the awk (or perl) backend.

Example: to select all rows that have valid data in column "a" or column "b" or column "c" you can

 vnl-filter 'a != "-" || b != "-" || c != "-"'

or

 vnl-filter --perl 'defined a || defined b || defined c'

As with the named expressions given to "-p" (described above), these are passed directly to awk, so anything that can be done with awk is supported here.

--eval expr

Instead of printing out all matching records and picked columns, just run the given chunk of awk (or perl). In this mode of operation, "vnl-filter" acts just like a glorified awk, that allows fields to be accessed by name instead of by number, as it would be in raw awk.

Since the expression may print anything or nothing at all, the output in this mode is not necessarily itself a valid vnlog stream. And no column-selecting arguments should be given, since they make no sense in this mode.

In awk the expr is a full set of pattern/action statements. So to print the sum of columns "a" and "b" in each row, and at the end, print the sum of all values in the "a" column

 vnl-filter --eval '{print a+b; suma += a} END {print suma}'

In perl the arbitrary expression fits in like this:

 while(<>) # read each line
 {
   next unless matches; # skip non-matching lines
   eval expression;     # evaluate the arbitrary expression
 }

--function|--sub

Evaluates the given expression as a function that can be used in other expressions. This is most useful when you want to print something that can't trivially be written as a simple expression. For instance:

 $ cat tst.vnl
 # s
 1-2
 3-4
 5-6
 $ < tst.vnl
   vnl-filter --function 'before(x) { sub("-.*","",x); return x }' \
              --function 'after(x)  { sub(".*-","",x); return x }' \
              -p 'b=before(s),a=after(s)'
 # b a
 1 2
 3 4
 5 6

See the CAVEATS section below if you're doing something sufficiently-complicated where you need this.

--[no]skipempty

Do [not] skip records where all fields are blank. By default we do skip all empty records; to include them, pass "--noskipempty"

--skipcomments

Don't output non-legend comments

--perl

By default all procesing is performed by "mawk", but if for whatever reason we want perl instead, pass "--perl". Both modes work, but "mawk" is noticeably faster. "--perl" could be useful because it is more powerful, which could be important since a number of things pass commandline strings directly to the underlying language (named expressions, matches expressions, "--eval" strings). Note that while variables in perl use sigils, column references should not use sigils. To print the sum of all values in column "a" you'd do this in awk

 vnl-filter --eval '{suma += a} END {print suma}'

and this in perl

 vnl-filter --perl --eval '{$suma += a} END {say $suma}'

The perl strings are evaluated without "use strict" or "use warnings" so I didn't have to declare $suma in the example.

--dumpexprs

Used for debugging. This spits out all the final awk (or perl) program we run for the given commandline options and given input. This is the final program, with the column references resolved to numeric indices, so one can figure out what went wrong.

--unbuffered

Flushes each line after each print. This makes sure each line is output as soon as it is available, which is crucial for realtime output and streaming plots.

--stream

Synonym for "--unbuffered"

This tool is very lax in its input validation (on purpose). As a result, columns with names like %CPU and "TIME+" do work (i.e. you can more or less feed in output from "top -b"). The downside is that shooting yourself in the foot is possible. This tradeoff is currently set to work well for my use cases, but I'd be interested in hearing other people's experiences. Potential pitfalls/unexpected behaviors:

  • When substituting column names I match either a word-nonword transition ("\b") or a whitespace-nonword transition. The word boundaries is what would be used 99% of the time. But the keys may have special characters in them, which don't work with "\b". This means that whitespace becomes important: "1+%CPU" will not be parsed as expected, which is correct since "+%CPU" is also a valid field name. But "1+ %CPU" will be parsed correctly, so if you have weird field names, put the whitespace into your expressions. It'll make them more readable anyway.
  • Strings passed to "-p" are split on "," except if the "," is inside balanced "()". This makes it possible to say things like "vnl-filter --function 'f(a,b) { ... }' -p 'c=f(a,b)'". This is probably the right behavior, although some questionable looking field names become potentially impossible: "f(a" and "b)" could otherwise be legal field names, but you're probably asking for trouble if you do that.
  • All column names are replaced in all eval strings without regard to context. The earlier example that reports the sum of values in a column: "vnl-filter --eval '{suma += a} END {print suma}'" will work fine if we do have a column named "a" and do <not> have a column named "suma". But will not do the right thing if any of those are violated. It's the user's responsibility to make sure we're talking about the right columns. The focus here was one-liners so hopefully nobody has so many columns, they can't keep track of all of them in their head. I don't see any way to resolve this without seriously impacting the scope of the tool, so I'm leaving this alone. Comments welcome.
  • Currently there're two modes: a pick/print mode and an "--eval" mode. Then there's also "--function", which adds bits of "--eval" to the pick/print mode, but it feels maybe insufficient. I don't yet have strong feelings about what this should become. Comments welcome

https://github.com/dkogan/vnlog/

Dima Kogan "<dima@secretsauce.net>"

Copyright 2016-2017 California Institute of Technology

Copyright 2017-2018 Dima Kogan "<dima@secretsauce.net>"

This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2.1 of the License, or (at your option) any later version.

2019-01-22