summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
Diffstat
-rw-r--r--binom_coeff.awk6
-rw-r--r--conv.awk1
-rw-r--r--cov.awk23
-rw-r--r--det.awk14
-rw-r--r--diff.awk10
-rw-r--r--diff1.awk10
-rw-r--r--em.awk7
-rw-r--r--exp_reg.awk16
-rw-r--r--fib.awk3
-rw-r--r--fir_window.awk (renamed from hamming.awk)11
-rw-r--r--fir_window_gaussian.awk (renamed from gaussian.awk)6
-rw-r--r--fir_window_hamming.awk41
-rw-r--r--fir_window_vonhann.awk35
-rw-r--r--kalman.awk30
-rw-r--r--kalman_init.awk26
-rw-r--r--lin_reg.awk17
-rw-r--r--lin_reg1.awk14
-rw-r--r--lin_reg2.awk14
-rw-r--r--lpf.awk24
-rw-r--r--ludcmp.awk8
-rw-r--r--lupd.awk13
-rw-r--r--mean.awk5
-rw-r--r--mean_avg.awk12
-rw-r--r--pi.awk5
-rw-r--r--quad_reg.awk16
-rw-r--r--sterling_approx.awk12
-rw-r--r--sum1.awk28
-rw-r--r--sum2.awk19
-rw-r--r--sum3.awk15
-rw-r--r--sum4.awk13
30 files changed, 325 insertions, 129 deletions
diff --git a/binom_coeff.awk b/binom_coeff.awk
index a19f464..6896671 100644
--- a/binom_coeff.awk
+++ b/binom_coeff.awk
@@ -1,7 +1,9 @@
-#! /usr/bin/awk -f
+#!/usr/bin/awk -f
### binomial coeffecient
# https://rosettacode.org/wiki/Evaluate_binomial_coefficients
+
+
function binom(n, k) {
b = 1
for (i=1; i<(k+1); i++) {
@@ -10,9 +12,11 @@ function binom(n, k) {
return b
}
+
BEGIN {
ARGV[1] ? N = ARGV[1] : N = 1
ARGV[2] ? K = ARGV[2] : K = 1
ARGV[3] ? OFMT = "%." ARGV[3] "g" : OFMT = "%g"
printf(OFMT ORS, binom(N, K))
}
+
diff --git a/conv.awk b/conv.awk
index b7e48eb..e46dcdf 100644
--- a/conv.awk
+++ b/conv.awk
@@ -3,6 +3,7 @@
### conv.awk
# [PoC] linear convolution (with hardcoded IR window).
+
BEGIN {
ARGV[1] ? X = ARGV[1] : X = 0
ARGV[2] ? OFMT = "%." ARGV[2] "g" : OFMT = "%g"
diff --git a/cov.awk b/cov.awk
index c06a447..92e92a8 100644
--- a/cov.awk
+++ b/cov.awk
@@ -3,6 +3,7 @@
### cov.awk
# online covariance algorithm
+
BEGIN {
OFS = FS
sign = "[+-]?"
@@ -12,18 +13,20 @@ BEGIN {
number = "^" sign "(" decimal "|" fraction ")" exponent "$"
}
+
+# column headers
NR == 1 {
- for (y=1; y<=NF; y++)
- ($y ~ number) ? header[y] = "col" y : header[y] = $y
+ for (n=1; n<=NF; n++)
+ ($n ~ number) ? header[n] = "col" n : header[n] = $n
}
+
NF > 0 {
- if (NF > nf_max)
- nf_max = NF
+ (NF > nf_max) ? nf_max = NF : nf_max = nf_max
- ### columns
+ # columns
for (y=1; y<=nf_max; y++) {
- ### rows
+ # rows
for (x=1; x<=nf_max; x++) {
if ($x !~ number || $x == header[x])
continue
@@ -39,18 +42,19 @@ NF > 0 {
}
}
+
END {
- ### column headers
+ # column headers
printf("cov")
for (y=1; y<=nf_max; y++) {
printf(OFS header[y])
}
printf(ORS)
- ### columns
+ # columns
for (y=1; y<=nf_max; y++) {
printf(header[y] OFS)
- ### rows
+ # rows
for (x=1; x<=nf_max; x++) {
printf(OFMT, cov_samp[x,y])
if (x < nf_max)
@@ -59,3 +63,4 @@ END {
printf(ORS)
}
}
+
diff --git a/det.awk b/det.awk
index b3e5c1f..e83a9ea 100644
--- a/det.awk
+++ b/det.awk
@@ -5,6 +5,7 @@
# input: square array as delimited text
# output: (scalar) determinant
+
BEGIN {
OFS = "\t"
sign = "[+-]?"
@@ -14,17 +15,21 @@ BEGIN {
number = "^" sign "(" decimal "|" fraction ")" exponent "$"
}
+
+# column headers
NR == 1 {
for (y=1; y<=NF; y++)
($y ~ number) ? header[y] = "col" y : header[y] = $y
}
+
+# read input data
NF > 0 {
(NF > nf_max) ? nf_max = NF : nf_max = nf_max
- ### columns
+ # columns
for (y=1; y<=nf_max; y++) {
- ### rows
+ # rows
if ($y !~ number) {
continue
}
@@ -36,12 +41,13 @@ NF > 0 {
}
}
+
END {
print length(row_sum), length(col_sum)
- ### columns
+ # columns
for (y=1; y<=nf_max; y++) {
if (y in col_sum) {
- ### rows
+ # rows
for (x=1; x<=NR; x++) {
if (x in row_sum) {
printf("[" OFMT "," OFMT "]" OFS OFMT OFS OFMT,
diff --git a/diff.awk b/diff.awk
index bf8b472..83fd0b5 100644
--- a/diff.awk
+++ b/diff.awk
@@ -3,9 +3,9 @@
### diff.awk
# print numerical diff along columns
+
BEGIN {
OFS = FS
- # OFMT = "%.9g"
sign = "[+-]?"
decimal = "[0-9]+[.]?[0-9]*"
fraction = "[.][0-9]*"
@@ -13,6 +13,8 @@ BEGIN {
number = "^" sign "(" decimal "|" fraction ")" exponent "$"
}
+
+# column headers
NR == 1 {
# orig data columns
for (n=1; n<=NF; n++) {
@@ -35,9 +37,9 @@ NR == 1 {
}
}
-NF {
- if (NF > nf_max)
- nf_max = NF
+
+NF > 0 {
+ (NF > nf_max) ? nf_max = NF : nf_max = nf_max
# data columns
for (n=1; n<=nf_max; n++) {
diff --git a/diff1.awk b/diff1.awk
index 7322e98..a6bf6ec 100644
--- a/diff1.awk
+++ b/diff1.awk
@@ -3,9 +3,9 @@
### diff1.awk
# print numerical diff along columns
+
BEGIN {
OFS = FS
- # OFMT = "%.9g"
sign = "[+-]?"
decimal = "[0-9]+[.]?[0-9]*"
fraction = "[.][0-9]*"
@@ -13,6 +13,8 @@ BEGIN {
number = "^" sign "(" decimal "|" fraction ")" exponent "$"
}
+
+# column headers
NR == 1 {
# orig data columns
for (n=1; n<=NF; n++) {
@@ -29,9 +31,9 @@ NR == 1 {
}
}
-NF {
- if (NF > nf_max)
- nf_max = NF
+
+NF > 0 {
+ (NF > nf_max) ? nf_max = NF : nf_max = nf_max
# data columns
for (n=1; n<=nf_max; n++) {
diff --git a/em.awk b/em.awk
index c956fa1..2cc6a74 100644
--- a/em.awk
+++ b/em.awk
@@ -1,9 +1,13 @@
#!/usr/bin/awk -f
+### em.awk
+# list of physical EM constants
+
+
BEGIN {
ARGV[1] ? OFMT = "%." ARGV[1] "g" : OFMT = "%.12g"
- ### constants
+ # constants
pi = 4.0*atan2(1,1) # rad
c0 = 299792458E0 # m/s, exact
mu0 = (4.0E-7)*pi # H/m = (T*m)/A = N/(A^2) = (kg*m)/(A*s)
@@ -11,3 +15,4 @@ BEGIN {
print pi, c0, mu0, epsilon0
}
+
diff --git a/exp_reg.awk b/exp_reg.awk
index f4ea12c..fac4b8f 100644
--- a/exp_reg.awk
+++ b/exp_reg.awk
@@ -3,6 +3,7 @@
### lin_reg.awk
# simple linear regression between columns
+
BEGIN {
OFS = ":"
sign = "[+-]?"
@@ -12,23 +13,25 @@ BEGIN {
number = "^" sign "(" decimal "|" fraction ")" exponent "$"
}
+
+# column headers
NR == 1 {
for (n=1; n<=NF; n++)
($n ~ number) ? header[n] = "col" n : header[n] = $n
}
-NF {
- if (NF > nf_max)
- nf_max = NF
- ### iterate over columns
+NF > 0 {
+ (NF > nf_max) ? nf_max = NF : nf_max = nf_max
+
+ # columns
for (y=1; y<=nf_max; y++) {
if ($y == header[n])
continue
if ($y ~ number) {
- ### mean
+ # mean
count[y] += 1
data[y] = log($y)
sum[y] += data[y]
@@ -39,7 +42,7 @@ NF {
sum_delta[y] += delta1[y]
sum_delta2[y] += delta0[y]*delta1[y]
- ### sample variance
+ # sample variance
#(count[y] > 1) ? var[y] = sum_delta2[y]/(count[y] - 1) : var[y] = ""
# x = row, y = col, trendline: y = A + Bx
@@ -92,6 +95,7 @@ NF {
}
}
+
END {
for (y=1; y<=nf_max; y++) {
for (x=1; x<=nf_max; x++) {
diff --git a/fib.awk b/fib.awk
index 7baef76..01cb8ce 100644
--- a/fib.awk
+++ b/fib.awk
@@ -1,6 +1,9 @@
#!/usr/bin/awk -f
### fib.awk
+# print nth Fibonacci number, based on R. Hamming's solution to a diff eqn
+# form of the sequence
+
BEGIN {
ARGV[1] ? n = ARGV[1] : n = 0
diff --git a/hamming.awk b/fir_window.awk
index acfde97..076e5df 100644
--- a/hamming.awk
+++ b/fir_window.awk
@@ -7,18 +7,18 @@
# 'a0' and 'a1' parameters of the raised cosine.
-### Gaussian
+# Gaussian window
function gaussian(n) {
# sigma <= 0.5
sigma = 0.4
- e[n] = (n)/(sigma*M)
+ en[n] = (n - M)/(sigma*M)
return exp(-0.5*e[n]*e[n])
}
-### Hamming, raised cosine
+# Hamming window (raised cosine)
function hamming(n) {
- ### optimal values for equal-ripple
+ # optimal values for equal-ripple
#a0 = 0.53836
#a1 = 0.46164
@@ -28,7 +28,7 @@ function hamming(n) {
}
-### vonHann, raised cosine
+# vonHann (raised cosine)
function vonhann(n) {
a0 = 0.5
a1 = 0.5
@@ -54,6 +54,5 @@ BEGIN {
print n + M, 1.0
}
}
-
}
diff --git a/gaussian.awk b/fir_window_gaussian.awk
index 402ea10..14ebc48 100644
--- a/gaussian.awk
+++ b/fir_window_gaussian.awk
@@ -1,11 +1,11 @@
#!/usr/bin/awk -f
-### gaussian.awk
+### fir_window_gaussian.awk
# generate a Gaussian window
# https://en.wikipedia.org/wiki/Window_function
-### Gaussian
+# Gaussian window
function gaussian(n) {
# sigma <= 0.5
sigma = 0.4
@@ -35,5 +35,5 @@ BEGIN {
print n, 1.0
}
}
-
}
+
diff --git a/fir_window_hamming.awk b/fir_window_hamming.awk
new file mode 100644
index 0000000..5660754
--- /dev/null
+++ b/fir_window_hamming.awk
@@ -0,0 +1,41 @@
+#!/usr/bin/awk -f
+
+### fir_window_hamming.awk
+# generate a Hamming window
+# R.W. Hamming, 'Digital Filters': H = '0.23 0.54 0.23'
+# https://en.wikipedia.org/wiki/Window_function provides a few values for the
+# 'a0' and 'a1' parameters of the raised cosine.
+
+
+# Hamming window (raised cosine)
+function hamming(n) {
+ # optimal values for equal-ripple
+ #a0 = 0.53836
+ #a1 = 0.46164
+
+ a0 = (25.0/46.0)
+ a1 = (21.0/46.0)
+ return a0 + a1*cos((pi*n)/M)
+}
+
+
+BEGIN {
+
+ ARGV[1] ? N = ARGV[1] : N = 0
+ ARGV[2] ? OFMT = "%." ARGV[2] "g" : OFMT = "%g"
+
+ # window interval goes from -M to M
+ M = 0.5*(N - 1)
+ pi = 4*atan2(1,1)
+
+ for (n=-M; n<=M; n++) {
+ if (N > 1 && M > 0) {
+ w[n] = hamming(n)
+ print n + M, w[n]/M
+ }
+ else {
+ print n + M, 1.0
+ }
+ }
+}
+
diff --git a/fir_window_vonhann.awk b/fir_window_vonhann.awk
new file mode 100644
index 0000000..3c831c7
--- /dev/null
+++ b/fir_window_vonhann.awk
@@ -0,0 +1,35 @@
+#!/usr/bin/awk -f
+
+### fir_window_vonhann.awk
+# https://en.wikipedia.org/wiki/Window_function provides a few values for the
+# 'a0' and 'a1' parameters of the raised cosine.
+
+
+# vonHann (raised cosine)
+function vonhann(n) {
+ a0 = 0.5
+ a1 = 0.5
+ return a0 + a1*cos((pi*n)/M)
+}
+
+
+BEGIN {
+
+ ARGV[1] ? N = ARGV[1] : N = 0
+ ARGV[2] ? OFMT = "%." ARGV[2] "g" : OFMT = "%g"
+
+ # window interval goes from -M to M
+ M = 0.5*(N - 1)
+ pi = 4*atan2(1,1)
+
+ for (n=-M; n<=M; n++) {
+ if (N > 1 && M > 0) {
+ w[n] = hamming(n)
+ print n + M, w[n]/M
+ }
+ else {
+ print n + M, 1.0
+ }
+ }
+}
+
diff --git a/kalman.awk b/kalman.awk
index f68cf81..7b6cb82 100644
--- a/kalman.awk
+++ b/kalman.awk
@@ -1,7 +1,7 @@
#!/usr/bin/awk -f
### kalman.awk
-# simple linear regression between columns
+
BEGIN {
sign = "[+-]?"
@@ -11,6 +11,8 @@ BEGIN {
number = "^" sign "(" decimal "|" fraction ")" exponent "$"
}
+
+# column headers
NR == 1 {
header_nf = NF
for (n=1; n<=NF; n++) {
@@ -18,31 +20,30 @@ NR == 1 {
}
}
-NF != 0 {
- #printf("\n%s: %s", NR, $0)
- if (NF > max_nf)
- max_nf = NF
- ### iterate over columns
- for (y=1; y<=max_nf; y++) {
+NF > 0 {
+ (NF > nf_max) ? nf_max = NF : nf_max = nf_max
+
+ # columns
+ for (y=1; y<=nf_max; y++) {
if ($y ~ number) {
- ### mean
+ # mean
count[y] += 1
sum[y] += $y
sum2[y] += $y*$y
mean[y] = sum[y]/count[y]
- ### difference from the mean
+ # difference from the mean
delta[y] = $y - mean[y]
sum_delta[y] += delta[y]
sum_delta2[y] += delta[y]*delta[y]
- ### sample variance
+ # sample variance
(count[y] - 1) ? var[y] = sum_delta2[y]/(count[y] - 1) : var[y] = 0
# x = row, y = col
- for (x=1; x<=max_nf; x++) {
+ for (x=1; x<=nf_max; x++) {
count[x,y] += 1
sum_xy[x,y] += $x*$y
sum_delta_xy[x,y] += delta[x]*delta[y]
@@ -72,7 +73,7 @@ NF != 0 {
b[x,y] = 1
}
- ### error estimate
+ # error estimate
err_den[x,y] = count[x,y]*(count[x,y] - 2)
if (count[x,y] > 2) {
err[x,y] = $y - (a[x,y] + b[x,y]*$x)
@@ -93,9 +94,10 @@ NF != 0 {
}
}
+
END {
- for (y=1; y<=max_nf; y++) {
- for (x=1; x<=max_nf; x++) {
+ for (y=1; y<=nf_max; y++) {
+ for (x=1; x<=nf_max; x++) {
if (x != y && r[x,y]) {
printf(OFMT OFS OFMT OFS "(%s)" OFS " = (" OFMT " +/- " OFMT ")(%s)" OFS " + (" OFMT " +/- " OFMT ")"ORS,
meas_cov[x,y], est_err_cov[x,y], header[y], b[x,y], b_err[x,y], header[x],
diff --git a/kalman_init.awk b/kalman_init.awk
index 57f7bbd..757d80b 100644
--- a/kalman_init.awk
+++ b/kalman_init.awk
@@ -1,7 +1,8 @@
#!/usr/bin/awk -f
-### kalman.awk
-# kalman csv experiemnt
+### kalman_init.awk
+# kalman filter csv experiemnt
+
BEGIN {
sign = "[+-]?"
@@ -11,6 +12,8 @@ BEGIN {
number = "^" sign "(" decimal "|" fraction ")" exponent "$"
}
+
+# column headers
NR == 1 {
header_nf = NF
for (n=1; n<=NF; n++) {
@@ -18,12 +21,12 @@ NR == 1 {
}
}
-NF != 0 {
+
+NF > 0 {
print ORS NR ":" OFS $0 ORS
- if (NF > max_nf)
- max_nf = NF
+ (NF > nf_max) ? nf_max = NF : nf_max = nf_max
- ### iterate over columns
+ # columns
for (n=1; n<=max_nf; n++) {
if ($n ~ number) {
count[n] = ($n ~ number) + (last[n] ~ number) + (last2[n] ~ number)
@@ -42,7 +45,7 @@ NF != 0 {
meas_err[n] = sqrt(var[n]/count[n])
}
- ### error covariance
+ # error covariance
for (m=1; m<=max_nf; m++) {
count[m,n]++
sum_xy[m,n] += 1
@@ -52,7 +55,7 @@ NF != 0 {
print meas_cov[m,n], est_err_last_cov[m,n]
}
- ### kalman gain
+ # kalman gain
if (count[n] <= 1) {
Gk[n] = 1
est_err[n] = sqrt(diff[n]*diff[n]) + 2.0*sqrt(diff2[n]*diff2[n])
@@ -61,17 +64,17 @@ NF != 0 {
else
Gk[n] = est_err_last[n]/(est_err_last[n] + meas_err[n])
- ### update estimate and estimated error
+ # update estimate and estimated error
est[n] = est_last[n] + Gk[n]*($n - est_last[n])
est_err[n] = (1.0 - Gk[n])*est_err_last[n]
- ### visual check
+ # visual check
printf "meas: " OFS OFMT OFS OFMT OFS OFMT, $n, meas_err[n], count[n]
printf "d1: " OFS OFMT, diff[n]
printf "d2: " OFS OFMT, diff2[n]
printf "est: " OFS OFMT OFS OFMT OFS OFMT, est[n], est_err[n], Gk[n]
- ### update previously remembered values
+ # update previously remembered values
last2[n] = last[n]
last[n] = $n
diff_last2[n] = diff_last[n]
@@ -86,3 +89,4 @@ NF != 0 {
}
}
}
+
diff --git a/lin_reg.awk b/lin_reg.awk
index 5a9df56..3006e8a 100644
--- a/lin_reg.awk
+++ b/lin_reg.awk
@@ -3,6 +3,7 @@
### lin_reg.awk
# simple linear regression between columns
+
BEGIN {
OFS = ":"
sign = "[+-]?"
@@ -12,23 +13,26 @@ BEGIN {
number = "^" sign "(" decimal "|" fraction ")" exponent "$"
}
+
+# column headers
NR == 1 {
for (n=1; n<=NF; n++)
($n ~ number) ? header[n] = "col" n : header[n] = $n
}
-NF {
- if (NF > nf_max)
- nf_max = NF
- ### iterate over columns
+# read input data
+NF > 0 {
+ (NF > nf_max) ? nf_max = NF : nf_max = nf_max
+
+ # columns
for (y=1; y<=nf_max; y++) {
if ($y == header[n])
continue
if ($y ~ number) {
- ### mean
+ # mean
count[y] += 1
sum[y] += $y
sum2[y] += $y*$y
@@ -38,7 +42,7 @@ NF {
sum_delta[y] += delta1[y]
sum_delta2[y] += delta0[y]*delta1[y]
- ### sample variance
+ # sample variance
#(count[y] > 1) ? var[y] = sum_delta2[y]/(count[y] - 1) : var[y] = ""
# x = row, y = col, trendline: y = A + Bx
@@ -91,6 +95,7 @@ NF {
}
}
+
END {
for (y=1; y<=nf_max; y++) {
for (x=1; x<=nf_max; x++) {
diff --git a/lin_reg1.awk b/lin_reg1.awk
index c657035..7786247 100644
--- a/lin_reg1.awk
+++ b/lin_reg1.awk
@@ -3,6 +3,7 @@
### lin_reg1.awk
# simple linear regression between columns
+
BEGIN {
sign = "[+-]?"
decimal = "[0-9]+[.]?[0-9]*"
@@ -11,22 +12,26 @@ BEGIN {
number = "^" sign "(" decimal "|" fraction ")" exponent "$"
}
+
+# column headers
NR == 1 {
for (n=1; n<=NF; n++)
($n ~ number) ? header[n] = "col" n : header[n] = $n
}
+
+# read input data
NF > 0 {
- if (NF > nf_max)
- nf_max = NF
+ (NF > nf_max) ? nf_max = NF : nf_max = nf_max
- ### iterate over columns
+ # columns
for (y=1; y<=nf_max; y++) {
if ($y == header[n])
continue
+
if ($y ~ number) {
- ### mean
+ # mean
count[y] += 1
sum[y] += $y
sum2[y] += $y*$y
@@ -62,6 +67,7 @@ NF > 0 {
}
}
+
END {
for (y=1; y<=nf_max; y++) {
for (x=1; x<=nf_max; x++) {
diff --git a/lin_reg2.awk b/lin_reg2.awk
index 450f0f0..d6097e3 100644
--- a/lin_reg2.awk
+++ b/lin_reg2.awk
@@ -3,6 +3,7 @@
### lin_reg2.awk
# simple linear regression between columns
+
BEGIN {
sign = "[+-]?"
decimal = "[0-9]+[.]?[0-9]*"
@@ -11,22 +12,26 @@ BEGIN {
number = "^" sign "(" decimal "|" fraction ")" exponent "$"
}
+
+# column headers
NR == 1 {
for (n=1; n<=NF; n++)
($n ~ number) ? header[n] = "col" n : header[n] = $n
}
+
+# read input data
NF > 0 {
- if (NF > nf_max)
- nf_max = NF
+ (NF > nf_max) ? nf_max = NF : nf_max = nf_max
- ### iterate over columns
+ # columns
for (y=1; y<=nf_max; y++) {
if ($y == header[n])
continue
+
if ($y ~ number) {
- ### mean
+ # mean
count[y] += 1
sum[y] += $y
sum2[y] += $y*$y
@@ -75,6 +80,7 @@ NF > 0 {
}
}
+
END {
for (y=1; y<=nf_max; y++) {
for (x=1; x<=nf_max; x++) {
diff --git a/lpf.awk b/lpf.awk
index aba92e1..638af3c 100644
--- a/lpf.awk
+++ b/lpf.awk
@@ -1,7 +1,8 @@
#!/usr/bin/awk -f
### lpf.awk
-# Low Pass Filter with Hardcoded FIR Window
+# low pass filter with hardcoded FIR window
+
BEGIN {
OFS = FS
@@ -20,19 +21,22 @@ BEGIN {
window_size = split(H, H_arr, "[ ]*")
}
+
+# column headers
NR == 1 {
for (y=1; y<=NF; y++)
($y ~ number) ? header[y] = "col" y : header[y] = $y
}
+
+# read input data
NF > 0 {
- if (NF > nf_max)
- nf_max = NF
+ (NF > nf_max) ? nf_max = NF : nf_max = nf_max
input_size = window_size
output_size = (input_size + window_size - 1)
- ### columns
+ # columns
for (y=1; y<=nf_max; y++) {
if ($y == header[y])
printf(header[y] OFS header[y] "_lpf")
@@ -75,15 +79,13 @@ NF > 0 {
printf(ORS)
}
-
-
-
}
+
END {
- ### rows
+ # rows
for (x=1; x<=window_size; x++) {
- ### columns
+ # columns
for (y=1; y<=nf_max; y++) {
# rotate input buffer
for (n=1; n<=input_size; n++) {
@@ -93,7 +95,7 @@ END {
#delete X_arr[y,input_size]
#input_size = length(X_arr)
#print length(X_arr)
-
+
Y[y] = 0
for (n=1; n<=window_size; n++) {
for (m=1; m<=input_size; m++) {
@@ -120,7 +122,7 @@ END {
printf(OFS)
else
printf(ORS)
- X_arr[input_size] = 0
+ X_arr[input_size] = 0
}
}
}
diff --git a/ludcmp.awk b/ludcmp.awk
index 6c7807a..8097263 100644
--- a/ludcmp.awk
+++ b/ludcmp.awk
@@ -11,6 +11,7 @@
# note, matrix indexing begins with 1 and uses i=row, j=col
# lu = L + U - I
+
# decomposition of A
function lu_decomp(A, n) {
@@ -43,6 +44,7 @@ function lu_decomp(A, n) {
}
}
+
# regex to identify strings that look like numbers
BEGIN {
OFS = FS
@@ -53,12 +55,14 @@ BEGIN {
number = "^" sign "(" decimal "|" fraction ")" exponent "$"
}
+
# column headers
NR == 1 {
for (n=1; n<=NF; n++)
($n ~ number) ? head[n] = "col" n : head[n] = $n
}
+
# read input data
NF {
@@ -74,17 +78,17 @@ NF {
print(matrix[count[n],n])
}
}
-
}
+
END {
printf(ORS)
print(NR, max_nf, size)
printf(ORS)
lu_decomp(matrix, size)
-
}
+
## find solution of Ly = b, for y
#for (i=0; i<n;, i++) {
# sum = 0
diff --git a/lupd.awk b/lupd.awk
index 39bf00c..08e1e30 100644
--- a/lupd.awk
+++ b/lupd.awk
@@ -9,6 +9,7 @@
# note, matrix indexing begins with 1 and uses i=row, j=col
# lu = L + U - I
+
# solve A[n,n] * x[n] = b[n]
function lu_solve(A, b, n) {
@@ -56,6 +57,7 @@ function lu_solve(A, b, n) {
}
+
# regex to identify strings that look like numbers
BEGIN {
OFS = FS
@@ -66,19 +68,21 @@ BEGIN {
number = "^" sign "(" decimal "|" fraction ")" exponent "$"
}
+
# column headers
NR == 1 {
for (y=1; y<=NF; y++)
($y ~ number) ? header[y] = "col" y : header[y] = $y
}
+
# read input data
NF > 0 {
(NF > nf_max) ? nf_max = NF : nf_max = nf_max
- ### columns
+ # columns
for (y=1; y<=nf_max; y++) {
- ### rows
+ # rows
if ($y !~ number) {
continue
}
@@ -90,12 +94,13 @@ NF > 0 {
}
}
+
END {
print length(row_sum), length(col_sum)
- ### columns
+ # columns
for (y=1; y<=nf_max; y++) {
if (y in col_sum) {
- ### rows
+ # rows
for (x=1; x<=NR; x++) {
if (x in row_sum) {
printf("[" OFMT "," OFMT "]" OFS OFMT OFS OFMT,
diff --git a/mean.awk b/mean.awk
index 069db4c..a57922b 100644
--- a/mean.awk
+++ b/mean.awk
@@ -3,6 +3,7 @@
### mean.awk
# calculate mean average of serialized input
+
BEGIN {
OFS = FS
#sign = "[+-]?"
@@ -12,8 +13,9 @@ BEGIN {
number = "^[+-]?([0-9]+[.]?[0-9]*|[.][0-9]*)([Ee][+-]?[0-9]+)?$"
}
+
# Welford's 'online' algorithm for variance
-NF {
+NF > 0 {
for (n=1; n<=NF; n++) {
if ($n ~ number) {
count += 1
@@ -29,6 +31,7 @@ NF {
}
}
+
END {
print "mean", "std_err", "std_dev", "range", "min", "max", "total", "count"
print mean, sqrt(var/count), sqrt(var), range, min, max, (mean*count), count
diff --git a/mean_avg.awk b/mean_avg.awk
index bef984d..05c93f1 100644
--- a/mean_avg.awk
+++ b/mean_avg.awk
@@ -5,6 +5,7 @@
# input: delimited data as text
# output: list of univariate summary stats
+
BEGIN {
OFS = FS
#sign = "[+-]?"
@@ -14,15 +15,19 @@ BEGIN {
number = "^[+-]?([0-9]+[.]?[0-9]*|[.][0-9]*)([Ee][+-]?[0-9]+)?$"
}
+
+# column headers
NR == 1 {
for (n=1; n<=NF; n++)
($n ~ number) ? header[n] = "col" n : header[n] = $n
}
+
# Welford's 'online' algorithm for variance
-NF {
- if (NF > max_nf)
- max_nf = NF
+NF > 0 {
+ (NF > nf_max) ? nf_max = NF : nf_max = nf_max
+
+ # columns
for (n=1; n <= NF; n++) {
if ($n ~ number) {
count[n] += 1
@@ -38,6 +43,7 @@ NF {
}
}
+
END {
print "col", "mean", "std_err", "std_dev", "range", "min", "max", "total", "count"
for (n=1; n<=max_nf; n++) {
diff --git a/pi.awk b/pi.awk
index 69eb36e..0a529f8 100644
--- a/pi.awk
+++ b/pi.awk
@@ -1,4 +1,4 @@
-#! /usr/bin/awk -f
+#!/usr/bin/awk -f
### pi.awk, https://en.wikipedia.org/wiki/Pi
# In 1706 John Machin used the Gregory–Leibniz series to produce an algorithm
@@ -11,11 +11,14 @@
# Ferguson–the best approximation achieved without the aid of a calculating
# device.
+
function pi() {
return 4*(4*atan2(1,5) - atan2(1,239))
}
+
BEGIN {
ARGV[1] ? OFMT = "%." ARGV[1] "g" : OFMT = "%g"
printf(OFMT ORS, pi())
}
+
diff --git a/quad_reg.awk b/quad_reg.awk
index edd0c6f..2682a0f 100644
--- a/quad_reg.awk
+++ b/quad_reg.awk
@@ -3,6 +3,7 @@
### quad_reg.awk
# quadratic regression along columns
+
BEGIN {
OFS = ":"
pi = 4.0*atan2(1.0, 1.0)
@@ -13,22 +14,26 @@ BEGIN {
number = "^" sign "(" decimal "|" fraction ")" exponent "$"
}
+
+# column headers
NR == 1 {
for (n=1; n<=NF; n++)
($n ~ number) ? header[n] = "col" n : header[n] = $n
}
+
+# read input data
NF > 0 {
- if (NF > nf_max)
- nf_max = NF
+ (NF > nf_max) ? nf_max = NF : nf_max = nf_max
- ### iterate over columns
+ # columns
for (y=1; y<=nf_max; y++) {
if ($y == header[n])
continue
+
if ($y ~ number) {
- ### mean
+ # mean
count[y] += 1
sum[y] += $y
sum2[y] += $y*$y
@@ -42,7 +47,7 @@ NF > 0 {
sum_delta[y] += delta1[y]
sum_delta2[y] += delta2[y]
- ### sample variance
+ # sample variance
#(count[y] > 1) ? var[y] = sum_delta2[y]/(count[y] - 1) : var[y] = ""
# x = row, y = col, trendline: y = A + Bx + Cx^2
@@ -115,6 +120,7 @@ NF > 0 {
}
}
+
END {
for (x=1; x<=nf_max; x++) {
for (y=1; y<=nf_max; y++) {
diff --git a/sterling_approx.awk b/sterling_approx.awk
index cb5f563..4b0741e 100644
--- a/sterling_approx.awk
+++ b/sterling_approx.awk
@@ -1,21 +1,20 @@
#!/usr/bin/awk -f
-
+### sterling_approx.awk
# https://en.wikipedia.org/wiki/Sterling_Approximation
# An alternative approximation for the Gamma function stated by Srinivasa
# Ramanujan (Ramanujan 1988) is
-# Gamma(1+x) ~= sqrt(pi)((x/e)^x)(8x^3 + 4x^2 + x + 1/30)^(1/6)
-# for x >= 0. The equivalent approximation for ln(n!) has an asymptotic error
-# of 1/(1400*n^3) ...
+# Gamma(1+x) ~= sqrt(pi)((x/e)^x)(8x^3 + 4x^2 + x + 1/30)^(1/6)
+# for x >= 0. The equivalent approximation for ln(n!) has an asymptotic
+# error of 1/(1400*n^3) ...
-### sterling_approx.awk
-# https://en.wikipedia.org/wiki/Stirling%27s_approximation
function pwr(x, p) {
return p ? exp(p*log(x)) : 1
}
+
BEGIN {
ARGV[1] ? n = ARGV[1] : n = 0
ARGV[2] ? OFMT = "%." ARGV[2] "g" : OFMT = "%g"
@@ -26,3 +25,4 @@ BEGIN {
}
printf(OFMT ORS, f)
}
+
diff --git a/sum1.awk b/sum1.awk
index 8f1419b..9266789 100644
--- a/sum1.awk
+++ b/sum1.awk
@@ -5,15 +5,35 @@
# output: sum of each column
# missing entries are treated as zeros
-BEGIN { OFS = FS }
-{
- if (NF > nf_max)
- nf_max = NF
+function isnum(n) {
+ sign = "[+-]?"
+ decimal = "[0-9]+[.]?[0-9]*"
+ fraction = "[.][0-9]+"
+ exponent = "([Ee]" sign "[0-9]+)?"
+ number = "^" sign "(" decimal "|" fraction ")" exponent "$"
+ return n ~ number
+}
+
+
+BEGIN {
+ OFS = FS
+}
+
+
+NR == 1 {
+ for (i=1; i<=NF; i++)
+ isnum($i) ? header[i] = "col" i : header[i] = $i
+}
+
+
+NF > 0 {
+ (NF > nf_max) ? nf_max = NF : nf_max = nf_max
for (i=1; i<=NF; i++)
sum[i] += $i
}
+
END {
for (i=1; i<=nf_max; i++) {
printf(OFMT, sum[i])
diff --git a/sum2.awk b/sum2.awk
index 694c047..4fcf3b6 100644
--- a/sum2.awk
+++ b/sum2.awk
@@ -3,18 +3,27 @@
### sum2.awk, print column sums
# check that each line has the same number of fields as line one
-BEGIN { OFS = FS }
-NR == 1 { nf_max = NF }
+BEGIN {
+ OFS = FS
+}
+
+
+NR == 1 {
+ for (i=1; i<=NF; i++)
+ isnum($i) ? header[i] = "col" i : header[i] = $i
+}
+
-{
+NF > 0 {
+ (NF > nf_max) ? nf_max = NF : nf_max = nf_max
for (i=1; i<=NF; i++)
sum[i] += $i
- if (NF != nf_max)
- print "line " NR " has " NF " entries, not " nf_max
}
+
END {
for (i=1; i<=nf_max; i++)
printf(OFMT "%s", sum[i], i < nf_max ? OFS : ORS)
}
+
diff --git a/sum3.awk b/sum3.awk
index 3e4661f..7f11a49 100644
--- a/sum3.awk
+++ b/sum3.awk
@@ -11,24 +11,29 @@ function isnum(n) {
}
-BEGIN { OFS = FS }
+BEGIN {
+ OFS = FS
+}
+
NR == 1 {
- nfld = NF
for (i=1; i<=NF; i++)
numcol[i] = isnum($i)
}
-{
+
+NF > 0 {
+ (NF > nf_max) ? nf_max = NF : nf_max = nf_max
for (i=1; i<=NF; i++)
if (numcol[i])
sum[i] += $i
}
+
END {
- for (i=1; i<=nfld; i++) {
+ for (i=1; i<=nf_max; i++) {
printf(numcol[i] ? sum[i] : "--")
- printf(i < nfld ? OFS : ORS)
+ printf(i < nf_max ? OFS : ORS)
}
}
diff --git a/sum4.awk b/sum4.awk
index aa3f044..3e61843 100644
--- a/sum4.awk
+++ b/sum4.awk
@@ -15,17 +15,19 @@ function isnum(n) {
}
-BEGIN { OFS = FS }
+BEGIN {
+ OFS = FS
+}
+
NR == 1 {
- nf_max = NF
for (i=1; i<=NF; i++)
isnum($i) ? header[i] = "col" i : header[i] = $i
}
-{
- if (NF > nf_max)
- nf_max = NF
+
+NF > 0 {
+ (NF > nf_max) ? nf_max = NF : nf_max = nf_max
for (i=1; i<=NF; i++) {
if ($i == header[i])
continue
@@ -36,6 +38,7 @@ NR == 1 {
}
}
+
END {
for (i=1; i<=nf_max; i++) {
printf((header[i]) ? header[i] OFS : OFS)
generated by cgit v1.2.3 (git 2.25.1) at 2026-07-25 14:54:49 +0000