AWK：帮助转换数据 table
AWK: Help on transforming data table
我有以下名为 in.txt 的文件：
2020-01-01 fruit banana 3.4
2020-03-02 alcohol smirnov 26.99
2020-03-10 fruit orange 4.20
2020-04-03 fruit orange 4.20
2021-09-01 alcohol beer 6.00
2021-08-03 fruit mango 6.99
2022-01-01 fruit orange 4.30
2022-03-04 alcohol beer 6.00
2022-03-03 alcohol beer 6.00
2022-04-01 fruit mango 7.20
我想转换文件，使其内容如下所示：
2020-01-01 2021-01-01 2022-01-01
-2020-12-31 -2021-12-31 -2022-12-31
fruit banana 3.40 0.00 0.00
orange 8.40 0.00 4.30
mango 0.00 6.99 7.20
Subt 11.80 6.99 11.50
alcohol beer 0.00 6.00 12.00
smirnov 26.99 0.00 0.00
Subt 26.99 6.00 12.00
Total 38.59 12.99 23.50
我已经开始编写以下脚本，但仍然不知道如何处理它。如何并排显示总计列。另一个问题是这只是虚拟数据。除了水果和酒精之外，我有许多不同的类别，为每个类别编写 if 语句和 for 循环似乎是错误的。另外，我如何才能只打印一次水果和酒精，而不是每次迭代第 3 列，并将日期范围置于顶部。非常感谢帮助。
#!/usr/bin/env bash
awk '
BEGIN{
FS=OFS="\t";
}
{
if ( ~ fruit && >= "2020-01-01" && <= "2020-12-31") {
a[]+=;
sa+=;
}
}
END {
PROCINFO["sorted_in"]="@ind_str_asc";
for (i in a) {
print "fruit", i, a[i]
}
}
' "${@:--}"
我相信下面这段 awk 代码是一个好的开始。剩下要做的只是一些清理和一些额外的求和代码。
BEGIN{
# how many divisions per year
n=1
# initialisation of some variables
tmax=0;tmin=999999; ymax=qmax=0;ymin=9999;qmin=99
}
# convert date to quarter,trim,half
{ y=+0; q=(substr(,6,7)+0)%n}
# compute min max time
(y*100+q < tmin) { ymin=y;qmin=q;tmin=y*100+q }
(y*100+q > tmax) { ymax=y;qmax=q;tmax=y*100+q }
# Create arrays that keep track of everything
# a : prices by year,q,category and element
# b : just a list of categories, eg fruit
# c : just a list of elements and the category it belongs to.
{ a[y,q,,]=; b[]; c[]= }
END{
# loop over categories (eg fruit)
for(i in b) {
# loop over elemnts
for(j in c) {
# exclude elements that do not belong to category
if (i!=c[j]) continue
s=i OFS j;
# loop over the time
for (y=ymin;y<=ymax;y++) {
for (q=0;q<n;++q) {
if (y*100+q < tmin) continue
if (y*100+q > tmax) continue
s=s OFS a[y,q,i,j]+0
}
}
print s
}
}
}
当前输出：
alcohol beer 0 6 6
alcohol smirnov 26.99 0 0
fruit orange 4.2 0 4.3
fruit mango 0 6.99 7.2
fruit banana 3.4 0 0
请您尝试以下操作：
#!/bin/bash
awk '
{
year = substr(, 1, 4) # extract year
if (from == "" || from > year) from = year # first (smallest) year
if (to == "" || to < year) to = year # last (largest) year
if ( in category == 0) {
category[] = # map item to category
list[] = list[] fs[] # csv of items
fs[] = "," # delimiter for csv
}
sum[,year] += # sum of the item in the year
subt[,year] += # sum of the category in the year
ttl[year] += # sum in the year
}
END {
format1 = "%-10s%-10s" # format for the left cells
format2 = "%-16s" # format for the header
format3 = "%-16.2f" # format for the amounts
# print upper header
printf(format1, "", "")
for (y = from; y <= to; y++) {
printf(format2, y "-01-01")
}
print ""
# print second header
printf(format1, "", "")
for (y = from; y <= to; y++) {
printf(format2, "-" y "-12-31")
}
print ""
for (cat in list) { # loop over the categories ("fruit" and "alcohol")
n = split(list[cat], item, ",") # split into items
for (i = 1; i <= n; i++) { # loop over the items
printf(format1, i == 1 ? cat : "", item[i])
for (y = from; y <= to; y++) { # loop over years
printf(format3, sum[item[i],y]) # append the sum of the year
}
print "" # finally break the line
}
print "" # insert blank line
printf(format1, "Subt", "")
for (y = from; y <= to; y++) {
printf(format3, subt[cat,y]) # append the subtotal
}
print "\n"
}
printf(format1, "Total", "")
for (y = from; y <= to; y++) {
printf(format3, ttl[y]) # append the total amount
}
print ""
}
' in.txt
使用提供的输入输出：
2020-01-01 2021-01-01 2022-01-01
-2020-12-31 -2021-12-31 -2022-12-31
alcohol smirnov 26.99 0.00 0.00
beer 0.00 6.00 12.00
Subt 26.99 6.00 12.00
fruit banana 3.40 0.00 0.00
orange 8.40 0.00 4.30
mango 0.00 6.99 7.20
Subt 11.80 6.99 11.50
Total 38.79 12.99 23.50
商品顺序与OP不一致请见谅
对数组的数组使用 GNU awk：
$ cat tst.awk
BEGIN { OFS="\t" }
{
sub(/-.*/,"",)
minYear = ( NR==1 || < minYear ? : minYear )
maxYear = ( NR==1 || > maxYear ? : maxYear )
items[][]
vals[][][] +=
typeTots[][] +=
yearTots[] +=
}
END {
printf "%s", OFS
for ( year=minYear; year<=maxYear; year++ ) {
printf "%s%s", OFS, year
}
print ""
for ( type in items ) {
itemCnt = 0
for ( item in items[type] ) {
printf "%s%s%s", (itemCnt++ ? "" : type), OFS, item
for ( year=minYear; year<=maxYear; year++ ) {
printf "%s%0.2f", OFS, vals[year][type][item]
}
print ""
}
printf "Subt%s", OFS
for ( year=minYear; year<=maxYear; year++ ) {
printf "%s%0.2f", OFS, typeTots[year][type]
}
print ORS
}
printf "Total%s", OFS
for ( year=minYear; year<=maxYear; year++ ) {
printf "%s%0.2f", OFS, yearTots[year]
}
print ""
}
$ awk -f tst.awk in.txt
2020 2021 2022
alcohol beer 0.00 6.00 12.00
smirnov 26.99 0.00 0.00
Subt 26.99 6.00 12.00
fruit orange 8.40 0.00 4.30
mango 0.00 6.99 7.20
banana 3.40 0.00 0.00
Subt 11.80 6.99 11.50
Total 38.79 12.99 23.50
或者如果您真的想要特定的日期范围，而不仅仅是 header 中的年份：
$ cat tst.awk
BEGIN { OFS="\t" }
{
sub(/-.*/,"",)
minYear = ( NR==1 || < minYear ? : minYear )
maxYear = ( NR==1 || > maxYear ? : maxYear )
items[][]
vals[][][] +=
typeTots[][] +=
yearTots[] +=
}
END {
printf "%s", OFS
for ( year=minYear; year<=maxYear; year++ ) {
printf "%s%s-01-01", OFS, year
}
print ""
printf "%s", OFS
for ( year=minYear; year<=maxYear; year++ ) {
printf "%s-%s-12-31", OFS, year
}
print ""
for ( type in items ) {
itemCnt = 0
for ( item in items[type] ) {
printf "%s%s%s", (itemCnt++ ? "" : type), OFS, item
for ( year=minYear; year<=maxYear; year++ ) {
printf "%s%0.2f", OFS, vals[year][type][item]
}
print ""
}
printf "Subt%s", OFS
for ( year=minYear; year<=maxYear; year++ ) {
printf "%s%0.2f", OFS, typeTots[year][type]
}
print ORS
}
printf "Total%s", OFS
for ( year=minYear; year<=maxYear; year++ ) {
printf "%s%0.2f", OFS, yearTots[year]
}
print ""
}
$ awk -f tst.awk in.txt | column -s$'\t' -t
2020-01-01 2021-01-01 2022-01-01
-2020-12-31 -2021-12-31 -2022-12-31
alcohol beer 0.00 6.00 12.00
smirnov 26.99 0.00 0.00
Subt 26.99 6.00 12.00
fruit orange 8.40 0.00 4.30
mango 0.00 6.99 7.20
banana 3.40 0.00 0.00
Subt 11.80 6.99 11.50
Total 38.79 12.99 23.50
我有以下名为 in.txt 的文件：
2020-01-01 fruit banana 3.4
2020-03-02 alcohol smirnov 26.99
2020-03-10 fruit orange 4.20
2020-04-03 fruit orange 4.20
2021-09-01 alcohol beer 6.00
2021-08-03 fruit mango 6.99
2022-01-01 fruit orange 4.30
2022-03-04 alcohol beer 6.00
2022-03-03 alcohol beer 6.00
2022-04-01 fruit mango 7.20
我想转换文件，使其内容如下所示：
2020-01-01 2021-01-01 2022-01-01
-2020-12-31 -2021-12-31 -2022-12-31
fruit banana 3.40 0.00 0.00
orange 8.40 0.00 4.30
mango 0.00 6.99 7.20
Subt 11.80 6.99 11.50
alcohol beer 0.00 6.00 12.00
smirnov 26.99 0.00 0.00
Subt 26.99 6.00 12.00
Total 38.59 12.99 23.50
我已经开始编写以下脚本，但仍然不知道如何处理它。如何并排显示总计列。另一个问题是这只是虚拟数据。除了水果和酒精之外，我有许多不同的类别，为每个类别编写 if 语句和 for 循环似乎是错误的。另外，我如何才能只打印一次水果和酒精，而不是每次迭代第 3 列，并将日期范围置于顶部。非常感谢帮助。
#!/usr/bin/env bash
awk '
BEGIN{
FS=OFS="\t";
}
{
if ( ~ fruit && >= "2020-01-01" && <= "2020-12-31") {
a[]+=;
sa+=;
}
}
END {
PROCINFO["sorted_in"]="@ind_str_asc";
for (i in a) {
print "fruit", i, a[i]
}
}
' "${@:--}"
我相信下面这段 awk 代码是一个好的开始。剩下要做的只是一些清理和一些额外的求和代码。
BEGIN{
# how many divisions per year
n=1
# initialisation of some variables
tmax=0;tmin=999999; ymax=qmax=0;ymin=9999;qmin=99
}
# convert date to quarter,trim,half
{ y=+0; q=(substr(,6,7)+0)%n}
# compute min max time
(y*100+q < tmin) { ymin=y;qmin=q;tmin=y*100+q }
(y*100+q > tmax) { ymax=y;qmax=q;tmax=y*100+q }
# Create arrays that keep track of everything
# a : prices by year,q,category and element
# b : just a list of categories, eg fruit
# c : just a list of elements and the category it belongs to.
{ a[y,q,,]=; b[]; c[]= }
END{
# loop over categories (eg fruit)
for(i in b) {
# loop over elemnts
for(j in c) {
# exclude elements that do not belong to category
if (i!=c[j]) continue
s=i OFS j;
# loop over the time
for (y=ymin;y<=ymax;y++) {
for (q=0;q<n;++q) {
if (y*100+q < tmin) continue
if (y*100+q > tmax) continue
s=s OFS a[y,q,i,j]+0
}
}
print s
}
}
}
当前输出：
alcohol beer 0 6 6
alcohol smirnov 26.99 0 0
fruit orange 4.2 0 4.3
fruit mango 0 6.99 7.2
fruit banana 3.4 0 0
请您尝试以下操作：
#!/bin/bash
awk '
{
year = substr(, 1, 4) # extract year
if (from == "" || from > year) from = year # first (smallest) year
if (to == "" || to < year) to = year # last (largest) year
if ( in category == 0) {
category[] = # map item to category
list[] = list[] fs[] # csv of items
fs[] = "," # delimiter for csv
}
sum[,year] += # sum of the item in the year
subt[,year] += # sum of the category in the year
ttl[year] += # sum in the year
}
END {
format1 = "%-10s%-10s" # format for the left cells
format2 = "%-16s" # format for the header
format3 = "%-16.2f" # format for the amounts
# print upper header
printf(format1, "", "")
for (y = from; y <= to; y++) {
printf(format2, y "-01-01")
}
print ""
# print second header
printf(format1, "", "")
for (y = from; y <= to; y++) {
printf(format2, "-" y "-12-31")
}
print ""
for (cat in list) { # loop over the categories ("fruit" and "alcohol")
n = split(list[cat], item, ",") # split into items
for (i = 1; i <= n; i++) { # loop over the items
printf(format1, i == 1 ? cat : "", item[i])
for (y = from; y <= to; y++) { # loop over years
printf(format3, sum[item[i],y]) # append the sum of the year
}
print "" # finally break the line
}
print "" # insert blank line
printf(format1, "Subt", "")
for (y = from; y <= to; y++) {
printf(format3, subt[cat,y]) # append the subtotal
}
print "\n"
}
printf(format1, "Total", "")
for (y = from; y <= to; y++) {
printf(format3, ttl[y]) # append the total amount
}
print ""
}
' in.txt
使用提供的输入输出：
2020-01-01 2021-01-01 2022-01-01
-2020-12-31 -2021-12-31 -2022-12-31
alcohol smirnov 26.99 0.00 0.00
beer 0.00 6.00 12.00
Subt 26.99 6.00 12.00
fruit banana 3.40 0.00 0.00
orange 8.40 0.00 4.30
mango 0.00 6.99 7.20
Subt 11.80 6.99 11.50
Total 38.79 12.99 23.50
商品顺序与OP不一致请见谅
对数组的数组使用 GNU awk：
$ cat tst.awk
BEGIN { OFS="\t" }
{
sub(/-.*/,"",)
minYear = ( NR==1 || < minYear ? : minYear )
maxYear = ( NR==1 || > maxYear ? : maxYear )
items[][]
vals[][][] +=
typeTots[][] +=
yearTots[] +=
}
END {
printf "%s", OFS
for ( year=minYear; year<=maxYear; year++ ) {
printf "%s%s", OFS, year
}
print ""
for ( type in items ) {
itemCnt = 0
for ( item in items[type] ) {
printf "%s%s%s", (itemCnt++ ? "" : type), OFS, item
for ( year=minYear; year<=maxYear; year++ ) {
printf "%s%0.2f", OFS, vals[year][type][item]
}
print ""
}
printf "Subt%s", OFS
for ( year=minYear; year<=maxYear; year++ ) {
printf "%s%0.2f", OFS, typeTots[year][type]
}
print ORS
}
printf "Total%s", OFS
for ( year=minYear; year<=maxYear; year++ ) {
printf "%s%0.2f", OFS, yearTots[year]
}
print ""
}
$ awk -f tst.awk in.txt
2020 2021 2022
alcohol beer 0.00 6.00 12.00
smirnov 26.99 0.00 0.00
Subt 26.99 6.00 12.00
fruit orange 8.40 0.00 4.30
mango 0.00 6.99 7.20
banana 3.40 0.00 0.00
Subt 11.80 6.99 11.50
Total 38.79 12.99 23.50
或者如果您真的想要特定的日期范围，而不仅仅是 header 中的年份：
$ cat tst.awk
BEGIN { OFS="\t" }
{
sub(/-.*/,"",)
minYear = ( NR==1 || < minYear ? : minYear )
maxYear = ( NR==1 || > maxYear ? : maxYear )
items[][]
vals[][][] +=
typeTots[][] +=
yearTots[] +=
}
END {
printf "%s", OFS
for ( year=minYear; year<=maxYear; year++ ) {
printf "%s%s-01-01", OFS, year
}
print ""
printf "%s", OFS
for ( year=minYear; year<=maxYear; year++ ) {
printf "%s-%s-12-31", OFS, year
}
print ""
for ( type in items ) {
itemCnt = 0
for ( item in items[type] ) {
printf "%s%s%s", (itemCnt++ ? "" : type), OFS, item
for ( year=minYear; year<=maxYear; year++ ) {
printf "%s%0.2f", OFS, vals[year][type][item]
}
print ""
}
printf "Subt%s", OFS
for ( year=minYear; year<=maxYear; year++ ) {
printf "%s%0.2f", OFS, typeTots[year][type]
}
print ORS
}
printf "Total%s", OFS
for ( year=minYear; year<=maxYear; year++ ) {
printf "%s%0.2f", OFS, yearTots[year]
}
print ""
}
$ awk -f tst.awk in.txt | column -s$'\t' -t
2020-01-01 2021-01-01 2022-01-01
-2020-12-31 -2021-12-31 -2022-12-31
alcohol beer 0.00 6.00 12.00
smirnov 26.99 0.00 0.00
Subt 26.99 6.00 12.00
fruit orange 8.40 0.00 4.30
mango 0.00 6.99 7.20
banana 3.40 0.00 0.00
Subt 11.80 6.99 11.50
Total 38.79 12.99 23.50