Wrong number of data blocks given by stats command - gnuplot

Question

I am using the following Gnuplot script with Gnuplot version 4.6 patchlevel 5 :
##### Prologue #####
clear # erases the current screen or output device
reset # all graph-related options take on their default values
###### Plot options #####
set style data lines
set surface
set dgrid3d 64,64 qnorm 2
set hidden3d
set ticslevel 0.8
set isosample 40,40
set view 60, 30, 1, 1
set contour base
###### Plot data #####
stats 'modele.out'
do for [i=0:int(STATS_blocks-1)] {
set key title 'density at t'.i
splot 'modele.out' every :::i::i using 2:3:5 notitle
pause 0.5
}
and the following file :
# Time is 0.000000000000000E+000
0.0000 0.0000 0.0000 0.9787 0.0213
0.0000 0.0000 0.3333 0.9077 0.0923
0.0000 0.0000 0.6667 0.9738 0.0262
0.0000 0.0000 1.0000 0.9673 0.0327
0.0000 0.3333 0.0000 0.9044 0.0956
0.0000 0.3333 0.3333 0.9246 0.0754
0.0000 0.3333 0.6667 0.9181 0.0819
0.0000 0.3333 1.0000 0.9089 0.0911
0.0000 0.6667 0.0000 0.9348 0.0652
0.0000 0.6667 0.3333 0.9372 0.0628
0.0000 0.6667 0.6667 0.9933 0.0067
0.0000 0.6667 1.0000 0.9273 0.0727
0.0000 1.0000 0.0000 0.9909 0.0091
0.0000 1.0000 0.3333 0.9771 0.0229
0.0000 1.0000 0.6667 0.9014 0.0986
0.0000 1.0000 1.0000 0.9312 0.0688
# Time is 50.0000000000000
50.0000 0.0000 0.0000 0.1036 0.8370
50.0000 0.0000 0.3333 0.1036 0.9093
50.0000 0.0000 0.6667 0.1031 0.9368
50.0000 0.0000 1.0000 0.1042 0.8378
50.0000 0.3333 0.0000 0.1034 0.9556
50.0000 0.3333 0.3333 0.1039 0.9127
50.0000 0.3333 0.6667 0.1041 0.9761
50.0000 0.3333 1.0000 0.1041 0.9587
50.0000 0.6667 0.0000 0.1033 0.9432
50.0000 0.6667 0.3333 0.1043 0.9503
50.0000 0.6667 0.6667 0.1087 0.5931
50.0000 0.6667 1.0000 0.1057 0.9579
50.0000 1.0000 0.0000 0.1044 0.8390
50.0000 1.0000 0.3333 0.1046 0.9101
50.0000 1.0000 0.6667 0.1062 0.9597
50.0000 1.0000 1.0000 0.1063 0.8494
# Time is 100.000000000000
100.0000 0.0000 0.0000 0.0997 0.8433
100.0000 0.0000 0.3333 0.0998 0.9123
100.0000 0.0000 0.6667 0.0995 0.9501
100.0000 0.0000 1.0000 0.0999 0.8442
100.0000 0.3333 0.0000 0.0999 0.9593
100.0000 0.3333 0.3333 0.1000 0.9157
100.0000 0.3333 0.6667 0.1000 0.9794
100.0000 0.3333 1.0000 0.1002 0.9612
100.0000 0.6667 0.0000 0.0997 0.9534
100.0000 0.6667 0.3333 0.1000 0.9542
100.0000 0.6667 0.6667 0.1001 0.6028
100.0000 0.6667 1.0000 0.1004 0.9584
100.0000 1.0000 0.0000 0.1000 0.8448
100.0000 1.0000 0.3333 0.1002 0.9143
100.0000 1.0000 0.6667 0.1005 0.9571
100.0000 1.0000 1.0000 0.1006 0.8490
I don't understand why the stats command returns that I have only 1 data block. That should be 3 in my opinion. Is the file badly formatted ?

Answer 1

stats gives you the number of indexable blocks in your data file. These blocks are separated by pairs of blank records (i.e. two blank lines).
If you did plot 'modele.out' index 0 you would find that it plotted all your data points as well, whereas index 1 would give you an error. There is only one (indexable) block in your data.
The solution
separate your blocks by two blank lines
change your splot command to splot 'modele.out' index i using 2:3:5 notitle
When you are using splot, a single blank line separates each row (or datablock, to use the term in the manual). This isn't the same thing as a block! In all other contexts (as far as I'm aware) there are two lines between each block (or indexable block to use the term in the manual).
update
As suggested by Christoph in the comments, if you wanted to keep your file in the same format and were sure that there were no blank lines at the end, you could change your loop to this:
do for [i=0:STATS_blank] {
and use your original splot line (with every, rather than index).