THD_zzprintf() long string truncation = the ZSS syndrome

AFNI Message Board
1

Hey Afni friends,

I’m running afni inside of a container right now, one of the commands I’m running is truncating the output for some unknown reason


3dinfo -verb  /mnt/150423/RESULTS/Stern/baseline_ListLength_EVENTS_censored/STATS_150423_REML.nii.gz | grep "Coef" | grep 'not5_LL5'

this returns


*+ WARNING: THD_zzprintf() long string truncation = the ZSS syndrome

and a truncated version of my output.
Its outputting 14 lines but my expected output is only around 20 lines.

AFNI version=AFNI_20.1.01 (Apr 14 2020) [64-bit]

Any ideas whats happening?

Thanks!

2

Hi-

This doesn’t answer your question directly, but would you get your parsing to work with the following?:


echo `3dinfo -label -sb_delim "\n"  /mnt/150423/RESULTS/Stern/baseline_ListLength_EVENTS_censored/STATS_150423_REML.nii.gz` | grep "Coef" | grep 'not5_LL5'

or, more succinctly:


echo `3dinfo -label -sb_delim "\n"  FILE` | grep "Coef" | grep 'not5_LL5'

–pt

3

It seems like sb_delim is using the literal characters of '' and ‘n’ so I altered your code a bit and got something to work:


3dinfo -label -sb_delim "," ${FILE} | tr ',' '\n' | grep 'Coef' | grep 'not5_LL5'

It’s a little round about but it should work as a workaround for know.

4

Howdy-

That’s why I had the


echo ` .... `

around the 3dinfo command-- I think that should still work if you include everything in that line?

Or is the problem that you want to wrap everything in ..., say, to save the output as a variable? In that case, you could do something like:


#tcsh 
set variable = `3dinfo -label  FILE | tr '|' '\n' | grep Coef`
# bash
variable=`3dinfo -label  FILE | tr '|' '\n' | grep Coef`

… which is basically what you had, but you don’t need to change the default subbrick delimiter, then.

–pt

5

Thanks for the quick replies,

Basically I’m trying to get the index of certain subbricks within an output
the back ticks seemed to mess with the output formatting a bit
I was able to get something working using the following:


3dinfo -subbrick_info ${input_file} | grep "${subbrick}" | grep "'${name}#" > $TFILE
idx=$(echo $(awk '{print $4}' $TFILE))
echo "${idx[@]//#}" > "${work_dir}/idx_${name}_${subbrick}.1D"

I would still be curious to know the reason for the THD_zzprintf() issue.

6

The THD_zzprintf() function merely has a limit on the maximum length. That could be increased if it is important.

If you are trying to get a single index, would -label2index be easier? You need to know the exact label for that.

3dinfo -label2index ‘aud#0_Coef’ stats.FT+tlrc

  • rick
7

Unfortunately I’m extracting multiple subbricks at once:


 -- At sub-brick #418 'not5_LL5#0_Coef' datum type is float:     -56.7884 to       84.4562
  -- At sub-brick #420 'not5_LL5#1_Coef' datum type is float:     -68.8189 to       78.3945
  -- At sub-brick #422 'not5_LL5#2_Coef' datum type is float:     -78.8084 to       100.778
  -- At sub-brick #424 'not5_LL5#3_Coef' datum type is float:     -72.1925 to       78.0441
  -- At sub-brick #426 'not5_LL5#4_Coef' datum type is float:     -80.5103 to       74.3215
  -- At sub-brick #428 'not5_LL5#5_Coef' datum type is float:     -83.7845 to       76.9669
  -- At sub-brick #430 'not5_LL5#6_Coef' datum type is float:     -69.4753 to       76.8984
  -- At sub-brick #432 'not5_LL5#7_Coef' datum type is float:      -67.924 to       61.0146
  -- At sub-brick #434 'not5_LL5#8_Coef' datum type is float:     -67.4012 to       76.7596
  -- At sub-brick #436 'not5_LL5#9_Coef' datum type is float:     -71.3718 to       54.9496
  -- At sub-brick #438 'not5_LL5#10_Coef' datum type is float:      -57.654 to       64.9789
  -- At sub-brick #440 'not5_LL5#11_Coef' datum type is float:     -70.3451 to       74.7157
  -- At sub-brick #442 'not5_LL5#12_Coef' datum type is float:      -71.217 to        90.233
  -- At sub-brick #444 'not5_LL5#13_Coef' datum type is float:     -71.6169 to       58.2551
  -- At sub-brick #446 'not5_LL5#14_Coef' datum type is float:     -56.2692 to       100.694
  -- At sub-brick #448 'not5_LL5#15_Coef' datum type is float:     -65.0389 to       76.3429
  -- At sub-brick #450 'not5_LL5#16_Coef' datum type is float:     -73.6157 to       81.1147
  -- At sub-brick #452 'not5_LL5#17_Coef' datum type is float:     -67.8461 to       80.9208
  -- At sub-brick #454 'not5_LL5#18_Coef' datum type is float:     -61.3783 to       66.8849
  -- At sub-brick #456 'not5_LL5#19_Coef' datum type is float:     -74.3717 to       86.5165
  -- At sub-brick #458 'not5_LL5#20_Coef' datum type is float:     -62.4366 to       61.9594

So i think I’ll have to do some funky clunky stuff. But this should work for now.
I was looking through the code here:
https://afni.nimh.nih.gov/pub/dist/src/debugtrace.c
is there a way for end users to set ZMAX?

8

Hi-

I guess there might be some spaces in the output when using ...? Not sure.

This worked for me to build a list of indices (in tcsh, but the syntax should transfer directly to bash):


#!/bin/tcsh

set FILE     = STATS_DSET
set idx_list = ( )   # to be populated

# make a list of all the labels
set all_labels = `3dinfo -label  "${FILE}" | tr '|' '\n' | grep Coef`

# go through the labels and populate list of indices
foreach label ( ${all_labels} )
    set idx_list = ( ${idx_list} `3dinfo -label2index "${label}" "${FILE}"` )
end

echo ""
echo "++ All my labels:"
echo "   ${all_labels}"
echo "++ My final index list:"
echo "   ${idx_list}"
9

Hi Rick,

We are running into this truncation error as well. You alluded to a way to "increase the maximum length if it is important’.
Could you please share this secret? :slight_smile:

Thank you!
Shruti

10

Hi Shruti,

Sorry for being slow. Could you provide some details regarding what you are seeing? Is it just from 3dinfo with a long history, for example?

Thanks,

  • rick