printf: zero pad after the decimal a given amount

[jidanni]

Why is there no way to tell printf to zero pad like the right column:
0.1 :0.100
0.05 :0.050
0.03 :0.030
0.025 :0.025
0.02 :0.020
0.015 :0.015
0.0125 :0.0125
0.01 :0.010
0.009 :0.009
0.00625:0.00625
0.005 :0.005
The challenge: Change only the "WHAT?" below to produce the right
column above. Thanks.
use constant S => 100000;
for ( 10000, 5000, 3000, 2500, 2000, 1500, 1250, 1000, 900, 625, 500 ) {
printf "%-7g:WHAT?\n", $_ / S, $_ / S;
}

 

[John W. Krahn]

Why is there no way to tell printf to zero pad like the right column:
0.1 :0.100
0.05 :0.050
0.03 :0.030
0.025 :0.025
0.02 :0.020
0.015 :0.015
0.0125 :0.0125
0.01 :0.010
0.009 :0.009
0.00625:0.00625
0.005 :0.005
The challenge: Change only the "WHAT?" below to produce the right
column above. Thanks.
use constant S => 100000;
for ( 10000, 5000, 3000, 2500, 2000, 1500, 1250, 1000, 900, 625, 500 ) {
printf "%-7g:WHAT?\n", $_ / S, $_ / S;
}

$ perl -le'
use constant S => 100000;
my $x;
format =
@<<<<<< : @.#####
$x, $x
..
for ( 10000, 5000, 3000, 2500, 2000, 1500, 1250, 1000, 900, 625, 500 ) {
$x = $_ / S;
write;
}
'
0.1 : 0.10000
0.05 : 0.05000
0.03 : 0.03000
0.025 : 0.02500
0.02 : 0.02000
0.015 : 0.01500
0.0125 : 0.01250
0.01 : 0.01000
0.009 : 0.00900
0.00625 : 0.00625
0.005 : 0.00500



John

 

[szr]

Why is there no way to tell printf to zero pad like the right column:
0.1 :0.100
0.05 :0.050
0.03 :0.030
0.025 :0.025
0.02 :0.020
0.015 :0.015
0.0125 :0.0125
0.01 :0.010
0.009 :0.009
0.00625:0.00625
0.005 :0.005
The challenge: Change only the "WHAT?" below to produce the right
column above. Thanks.
use constant S => 100000;
for ( 10000, 5000, 3000, 2500, 2000, 1500, 1250, 1000, 900, 625, 500
) { printf "%-7g:WHAT?\n", $_ / S, $_ / S;
}

use constant S => 100000;
for ( 10000, 5000, 3000, 2500, 2000, 1500, 1250, 1000, 900, 625, 500 ) {
printf "%-7g:%01d.%3.3s%s\n", $_ / S, int $_ / S,
sprintf("%05d", $_),
map { $_ ? $_ : '' } ($_ % 100) =~ m!^(\d+?)0*$!;
}

__OUTPUT__
0.1 :0.100
0.05 :0.050
0.03 :0.030
0.025 :0.025
0.02 :0.020
0.015 :0.015
0.0125 :0.0125
0.01 :0.010
0.009 :0.009
0.00625:0.00625
0.005 :0.005

 

[Paul Lalli]

Why is there no way to tell printf to zero pad like the right column:

Why do you assume that because you don't know the way, there is no
way?

0.1    :0.100
0.05   :0.050
0.03   :0.030
0.025  :0.025
0.02   :0.020
0.015  :0.015
0.0125 :0.0125
0.01   :0.010
0.009  :0.009
0.00625:0.00625
0.005  :0.005
The challenge: Change only the "WHAT?" below to produce the right
column above. Thanks.
use constant S => 100000;
for ( 10000, 5000, 3000, 2500, 2000, 1500, 1250, 1000, 900, 625, 500 ) {
    printf "%-7g:WHAT?\n", $_ / S, $_ / S;

%.03f

$ perl -e'printf("%.03f\n", .1)'
0.100

Paul Lalli

 

[szr]

Why do you assume that because you don't know the way, there is no
way?


%.03f

$ perl -e'printf("%.03f\n", .1)'
0.100

Paul Lalli

Actually that truncates to 3 decimal places, which isn't what the op
required:

$ perl -e'printf("%.03f\n", .00625)'
0.006


See my other post for a working solution.

 

[Dr.Ruud]

(e-mail address removed) schreef:

Why is there no way to tell printf to zero pad like the right column:
0.1 :0.100
0.05 :0.050
0.03 :0.030
0.025 :0.025
0.02 :0.020
0.015 :0.015
0.0125 :0.0125
0.01 :0.010
0.009 :0.009
0.00625:0.00625
0.005 :0.005

$ perl -wle'
print "".reverse sprintf "%05.1f", "".reverse sprintf "%f", $_
for qw/.1 .05 .03 .025 .02 .015 .0125 .01 .009 .00625 .005
1.987654321E1/
'
0.100
0.050
0.030
0.025
0.020
0.015
0.0125
0.010
0.009
0.00625
0.005
9.876543

 

[xhoster]

Why is there no way to tell printf to zero pad like the right column:

One reason is that what you want is ill-defined. If we are going to tweak
sprintf to make it suit our personal preferences, I'd rather see a
conversion character that behaved just like %f if given a good number, but
returned the empty string if given either an empty string or undef (rather
than converting it to zero and then applying %f to the zero.)

0.1 :0.100
0.05 :0.050
0.03 :0.030
0.025 :0.025
0.02 :0.020
0.015 :0.015

Apparently you want to preserve non-zero digits even if that means going
beyond 3 digits right of the decimal. But why did you stop at 4?

0.014999999999999999444888

0.0125 :0.0125

0.0125000000000000006938893

How many consecutive zeros or nines are needed before you decide there are
enough to ignore what is the right of them?

Xho

--
-------------------- http://NewsReader.Com/ --------------------
The costs of publication of this article were defrayed in part by the
payment of page charges. This article must therefore be hereby marked
advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate
this fact.

 

[Ben Morrow]

Quoth (e-mail address removed):

One reason is that what you want is ill-defined. If we are going to tweak
sprintf to make it suit our personal preferences, I'd rather see a
conversion character that behaved just like %f if given a good number, but
returned the empty string if given either an empty string or undef (rather
than converting it to zero and then applying %f to the zero.)


Apparently you want to preserve non-zero digits even if that means going
beyond 3 digits right of the decimal. But why did you stop at 4?

0.014999999999999999444888



0.0125000000000000006938893

How many consecutive zeros or nines are needed before you decide there are
enough to ignore what is the right of them?

It appears to me the OP wants either 3 s.f. after the point or 3 places,
whichever comes out shorter. Something like

sub fmt {
return
map /(\d*\.\d{3}\d*?)0*$/,
map /(\d*\.0*[1-9]\d\d)/,
map { sprintf "%.308f", $_ }
@_;
}

appears to work, but it's hardly pretty . The 308 is the number of
places required to represent DBL_MIN with 53-bit doubles; if your perl
is using 64-bit long doubles you will need 4932 instead.

Ben

 

[Ted Zlatanov]

BM> It appears to me the OP wants either 3 s.f. after the point or 3 places,
BM> whichever comes out shorter. Something like

BM> sub fmt {
BM> return
BM> map /(\d*\.\d{3}\d*?)0*$/,
BM> map /(\d*\.0*[1-9]\d\d)/,
BM> map { sprintf "%.308f", $_ }
BM> @_;
BM> }

BM> appears to work, but it's hardly pretty . The 308 is the number of
BM> places required to represent DBL_MIN with 53-bit doubles; if your perl
BM> is using 64-bit long doubles you will need 4932 instead.

Is there any harm in always using 4932? I would guess not, except maybe
for wasted CPU cycles.

Ted

 

[Dr.Ruud]

Frank Seitz schreef:

Dr.Ruud:

$ perl -wle'
print "".reverse sprintf "%05.1f", "".reverse sprintf "%f", $_
for qw/.1 .05 .03 .025 .02 .015 .0125 .01 .009 .00625 .005
1.987654321E1/
'
0.100
[...]
9.876543

And how do you deal with negative numbers and numbers >= 10?

Hey! There were only numbers in [0..1>, so I already extended it to
[0..10>.

Oh well,

perl -wle'
$n=length(int abs),
print+($_<0?"-":"").reverse sprintf"%0*.*f",$n+4,$n,
"".reverse sprintf"%f",abs
for qw/0 .1 .05 .03 .025 .02 .015 .0125 .01 .009 .00625 .005
1234567.89 -9876543.21/
'
0.000
0.100
0.050
0.030
0.025
0.020
0.015
0.0125
0.010
0.009
0.00625
0.005
1234567.890
-9876543.210