Selection based on segment header value -Zval

Q&A
1

Hello all,

is there a way to do selection from a multi segment file based on segment header values?
e.g. i have such a file and want to select segments with -Z > 0.3 (hope it is not too trivial !!!)

>-Z.390000                                                                      
 38.01660 38.21025  0.50000
 38.12926 38.22511  0.50000
 38.12893 38.22655  9.50000
 38.01627 38.21169  9.50000
>-Z.005270                                                                      
 38.00179 38.20296  0.58000
 37.96190 38.11862  0.58000
 37.98032 38.11319  9.42000
 38.02023 38.19752  9.42000
>-Z.037600                                                                      
 37.26096 38.01535  0.58000
 37.36783 37.98445  0.58000
 37.37458 37.99904  9.42000
 37.26769 38.02994  9.42000
>-Z.656000                                                                      
 37.20471 38.01427  0.50000
 37.31859 38.01423  0.50000
 37.31859 38.01567  9.50000
 37.20471 38.01571  9.50000
2

I would probably preprocess through a small perl program
something on the lines of:

$cap=0;
while(<>){
m/^>-Z(.*)/ and $cap=($1 > 0.3);
print $_ if($cap)
}

and then pipe to your chosen gmt module

3

I think not, but seems a reasonable feature request for gmt select -Z to accept a new modifier +h to pick up the z-value from the header instead of the third column(and hence it applies to each record in the segment).

4

This is also doable with csplit(1), I think.

Maybe a bit too much, but a fun exercise in bash.

# split every segment into individual files
$ csplit --prefix=split_ --elide-empty-files f "/>-Z/0" "{*}"
123
123
123
123

# get the name of files containing Z.3 (i.e. 0.3)
$ grep -l "Z\.3" split_0*
split_00
5

Very fragile because I didn’t care to parse the header and convert to floats

julia> D = gmtread("zvals_.dat")
Vector{GMTdataset{Float64, 2}} with 4 segments
Show first segment. To see other segments just type its element number. E.g. D[7]

BoundingBox: [38.01627, 38.12926, 38.21025, 38.22655, 0.5, 9.5]
Header: -Z.390000
4×3 GMTdataset{Float64, 2}
 Row │   col.1    col.2    col.3
     │ Float64  Float64  Float64
─────┼───────────────────────────
   1 │ 38.0166  38.2103      0.5
   2 │ 38.1293  38.2251      0.5
   3 │ 38.1289  38.2266      9.5
   4 │ 38.0163  38.2117      9.5

julia> ind = [D[k].header > "-Z.3" for k=1:length(D)]
4-element Vector{Bool}:
 1
 0
 0
 1

julia> D03 = D[ind]
Vector{GMTdataset{Float64, 2}} with 2 segments
Show first segment. To see other segments just type its element number. E.g. D[7]

BoundingBox: [38.01627, 38.12926, 38.21025, 38.22655, 0.5, 9.5]
Header: -Z.390000
4×3 GMTdataset{Float64, 2}
 Row │   col.1    col.2    col.3
     │ Float64  Float64  Float64
─────┼───────────────────────────
   1 │ 38.0166  38.2103      0.5
   2 │ 38.1293  38.2251      0.5
   3 │ 38.1289  38.2266      9.5
   4 │ 38.0163  38.2117      9.5
6

I cannot see csplit working for things like get segments where -Zval is between 4 and 6.7. Hard to beat

gmt select oldfile.txt -Z4/6.7+h > those only.txt

even in Julia or Python.

7

But there is no +h modifier in gmtselect -Z! And -Z operates on 3rd column while the question was on how operate on header encoded option.

8

How was the GMT vector file with the -Z values created? It might be easier to change the code that created the file in the first place.

9

Hello all,
THANK you for all suggestions !! i finally solved it using a small external code. Although such a code is easy to write (Maybe!) it would be nice for GMT to have such a feature included.

10

yes it would make sense but file was given to me !!

11

I know there is no +h. See earlier about me proposing it.

12

Such a filter option would be a great feature to have!

13

For those who wish to try this feature, check out this branch, build and explore: https://github.com/GenericMappingTools/gmt/pull/7309

14

Now approved and merged into the master branch so should be easy to get to for @chhei-s and @esokos to try out!