AWK：在 awk table 转换脚本中求和有问题
AWK: sums problematic in awk table transform script
我有以下文件：
cat st_in.txt
2015-01-01 2 A FI
2015-02-03 4 B VI
2015-03-01 6 A FI
2015-01-08 -4 C VE
2016-01-05 -3 B VE
2016-02-03 -1 D FE
2016-04-01 -2 B FE
2016-06-13 -5 D VE
2017-01-01 2 A VI
2017-02-03 3 A VI
2017-02-04 8 C FI
2017-01-05 -1 B FE
并想这样改造：
2015 2016 2017
A 8.00 0.00 0.00
C 0.00 0.00 8.00
sumFI 8.00 0.00 8.00
A 0.00 0.00 5.00
B 4.00 0.00 0.00
sumVI 4.00 0.00 5.00
sumI 12.00 0.00 13.00
B 0.00 -2.00 -1.00
D 0.00 -1.00 0.00
sumFE 0.00 -3.00 -1.00
B 0.00 -3.00 0.00
C -4.00 0.00 0.00
D 0.00 -5.00 0.00
sumVE -4.00 -8.00 0.00
sumE -4.00 -11.00 -1.00
net 8.00 -11.00 12.00
为此，我编写了以下脚本：
#!/usr/bin/env bash
awk '
BEGIN {
OFS = "\t"
}
{
yr = substr(, 1, 4)
sub(/-.*/, "", )
minYr = (NR == 1 || < minYr ? : minYr)
maxYr = (NR == 1 || > maxYr ? : maxYr)
H[][]
W[yr][][] +=
yT[] +=
val[][] +=
}
END {
for (yr = minYr; yr <= maxYr; yr++) {
printf "%s%s", OFS, yr
}
print ""
print ""
for (cT in H) {
for (c in H[cT]) {
printf c, OFS
for (yr = minYr; yr <= maxYr; yr++) {
printf "%s%0.2f", OFS, W[yr][cT][c]
}
print ""
}
printf "sum" cT, OFS
for (yr = minYr; yr <= maxYr; yr++) {
printf "%s%0.2f", OFS, val[yr][cT]
}
print ""
print ""
if (cT == "VI") {
printf "sumI", OFS
for (yr = minYr; yr <= maxYr; yr++) {
printf "%s%0.2f", OFS, W[yr][cT][c]
}
} else if (cT == "VE") {
printf "sumE", OFS
for (yr = minYr; yr <= maxYr; yr++) {
printf "%s%0.2f", OFS, W[yr][cT][c]
}
}
print ORS
}
printf "net"
for (yr = minYr; yr <= maxYr; yr++) {
printf "%s%0.2f", OFS, yT[yr]
}
print ""
}
' "${@:--}"
打印以下内容：
./trans1 st_in.txt
2015 2016 2017
A 8.00 0.00 0.00
C 0.00 0.00 8.00
sumFI 8.00 0.00 8.00
A 0.00 0.00 5.00
B 4.00 0.00 0.00
sumVI 4.00 0.00 5.00
sumI 4.00 0.00 0.00
B 0.00 -2.00 -1.00
D 0.00 -1.00 0.00
sumFE 0.00 -3.00 -1.00
B 0.00 -3.00 0.00
C -4.00 0.00 0.00
D 0.00 -5.00 0.00
sumVE -4.00 -8.00 0.00
sumE 0.00 -5.00 0.00
net 8.00 -11.00 12.00
我不担心这里的格式。控制台打印输出类似于目标。但我正在努力的是获得 sumI (sumFI + sumVI) 和 sumE(sumFE + sumVE) 的正确总和。有人可以帮忙吗？
Ed Morton 编辑以提供有意义的数据类型和变量名称，假设 OP 当前数据是 Date Amount Item Type 的列类型，就像在 :
中一样
$ cat trans1
#!/usr/bin/env bash
awk '
BEGIN {
OFS = "\t"
}
{
date =
amount =
item =
type =
year = substr(date, 1, 4)
minYear = (NR == 1 || year < minYear ? year : minYear)
maxYear = (NR == 1 || year > maxYear ? year : maxYear)
types_items[type][item]
yearsTypesItems2amounts[year][type][item] +=
years2amounts[year] +=
yearsTypes2amounts[year][type] +=
}
END {
for (year = minYear; year <= maxYear; year++) {
printf "%s%s", OFS, year
}
print ""
print ""
for (type in types_items) {
for (item in types_items[type]) {
printf item, OFS
for (year = minYear; year <= maxYear; year++) {
printf "%s%0.2f", OFS, yearsTypesItems2amounts[year][type][item]
}
print ""
}
printf "sum" type, OFS
for (year = minYear; year <= maxYear; year++) {
printf "%s%0.2f", OFS, yearsTypes2amounts[year][type]
}
print ""
print ""
if (type == "VI") {
printf "sumI", OFS
for (year = minYear; year <= maxYear; year++) {
printf "%s%0.2f", OFS, yearsTypesItems2amounts[year][type][item] # <--- NOTE: item unset here
}
} else if (type == "VE") {
printf "sumE", OFS
for (year = minYear; year <= maxYear; year++) {
printf "%s%0.2f", OFS, yearsTypesItems2amounts[year][type][item] # <--- NOTE: item unset here
}
}
print ORS
}
printf "net"
for (year = minYear; year <= maxYear; year++) {
printf "%s%0.2f", OFS, years2amounts[year]
}
print ""
}
' "${@:--}"
请参阅上面的“注意”，通过重命名练习可以明显看出某些错误所在（看起来您使用了错误的数组，应该使用 yearsTypes2amounts[year][type] 而不是 yearsTypesItems2amounts[year][type][item]).
继续对数组的数组使用 GNU awk，并假设 OP 当前数据是 Date Amount Item Type 的列类型，就像在 :
中一样
$ cat trans1
#!/usr/bin/env bash
awk '
BEGIN {
OFS = "\t"
}
{
date =
amount =
item =
type =
year = substr(date, 1, 4)
minYear = (NR == 1 || year < minYear ? year : minYear)
maxYear = (NR == 1 || year > maxYear ? year : maxYear)
cat = substr(type,2)
ctiys2amounts[cat][type][item][year] += amount
}
END {
for (year = minYear; year <= maxYear; year++) {
printf "%s%s", OFS, year
}
print ORS
for (cat in ctiys2amounts) {
delete catSum
for (type in ctiys2amounts[cat]) {
delete typeSum
for (item in ctiys2amounts[cat][type]) {
printf "%s", item
for (year = minYear; year <= maxYear; year++) {
amount = ctiys2amounts[cat][type][item][year]
printf "%s%0.2f", OFS, amount
typeSum[year] += amount
}
print ""
}
printf "sum%s", type
for (year = minYear; year <= maxYear; year++) {
printf "%s%0.2f", OFS, typeSum[year]
catSum[year] += typeSum[year]
}
print ORS
}
printf "sum%s", cat
for (year = minYear; year <= maxYear; year++) {
printf "%s%0.2f", OFS, catSum[year]
yearSum[year] += catSum[year]
}
print ORS
}
printf "net"
for (year = minYear; year <= maxYear; year++) {
printf "%s%0.2f", OFS, yearSum[year]
}
print ""
}
' "${@:--}"
$ ./trans1 st_in.txt
2015 2016 2017
B 0.00 -2.00 -1.00
D 0.00 -1.00 0.00
sumFE 0.00 -3.00 -1.00
B 0.00 -3.00 0.00
C -4.00 0.00 0.00
D 0.00 -5.00 0.00
sumVE -4.00 -8.00 0.00
sumE -4.00 -11.00 -1.00
A 8.00 0.00 0.00
C 0.00 0.00 8.00
sumFI 8.00 0.00 8.00
A 0.00 0.00 5.00
B 4.00 0.00 0.00
sumVI 4.00 0.00 5.00
sumI 12.00 0.00 13.00
net 8.00 -11.00 12.00
我有以下文件：
cat st_in.txt
2015-01-01 2 A FI
2015-02-03 4 B VI
2015-03-01 6 A FI
2015-01-08 -4 C VE
2016-01-05 -3 B VE
2016-02-03 -1 D FE
2016-04-01 -2 B FE
2016-06-13 -5 D VE
2017-01-01 2 A VI
2017-02-03 3 A VI
2017-02-04 8 C FI
2017-01-05 -1 B FE
并想这样改造：
2015 2016 2017
A 8.00 0.00 0.00
C 0.00 0.00 8.00
sumFI 8.00 0.00 8.00
A 0.00 0.00 5.00
B 4.00 0.00 0.00
sumVI 4.00 0.00 5.00
sumI 12.00 0.00 13.00
B 0.00 -2.00 -1.00
D 0.00 -1.00 0.00
sumFE 0.00 -3.00 -1.00
B 0.00 -3.00 0.00
C -4.00 0.00 0.00
D 0.00 -5.00 0.00
sumVE -4.00 -8.00 0.00
sumE -4.00 -11.00 -1.00
net 8.00 -11.00 12.00
为此，我编写了以下脚本：
#!/usr/bin/env bash
awk '
BEGIN {
OFS = "\t"
}
{
yr = substr(, 1, 4)
sub(/-.*/, "", )
minYr = (NR == 1 || < minYr ? : minYr)
maxYr = (NR == 1 || > maxYr ? : maxYr)
H[][]
W[yr][][] +=
yT[] +=
val[][] +=
}
END {
for (yr = minYr; yr <= maxYr; yr++) {
printf "%s%s", OFS, yr
}
print ""
print ""
for (cT in H) {
for (c in H[cT]) {
printf c, OFS
for (yr = minYr; yr <= maxYr; yr++) {
printf "%s%0.2f", OFS, W[yr][cT][c]
}
print ""
}
printf "sum" cT, OFS
for (yr = minYr; yr <= maxYr; yr++) {
printf "%s%0.2f", OFS, val[yr][cT]
}
print ""
print ""
if (cT == "VI") {
printf "sumI", OFS
for (yr = minYr; yr <= maxYr; yr++) {
printf "%s%0.2f", OFS, W[yr][cT][c]
}
} else if (cT == "VE") {
printf "sumE", OFS
for (yr = minYr; yr <= maxYr; yr++) {
printf "%s%0.2f", OFS, W[yr][cT][c]
}
}
print ORS
}
printf "net"
for (yr = minYr; yr <= maxYr; yr++) {
printf "%s%0.2f", OFS, yT[yr]
}
print ""
}
' "${@:--}"
打印以下内容：
./trans1 st_in.txt
2015 2016 2017
A 8.00 0.00 0.00
C 0.00 0.00 8.00
sumFI 8.00 0.00 8.00
A 0.00 0.00 5.00
B 4.00 0.00 0.00
sumVI 4.00 0.00 5.00
sumI 4.00 0.00 0.00
B 0.00 -2.00 -1.00
D 0.00 -1.00 0.00
sumFE 0.00 -3.00 -1.00
B 0.00 -3.00 0.00
C -4.00 0.00 0.00
D 0.00 -5.00 0.00
sumVE -4.00 -8.00 0.00
sumE 0.00 -5.00 0.00
net 8.00 -11.00 12.00
我不担心这里的格式。控制台打印输出类似于目标。但我正在努力的是获得 sumI (sumFI + sumVI) 和 sumE(sumFE + sumVE) 的正确总和。有人可以帮忙吗？
Ed Morton 编辑以提供有意义的数据类型和变量名称，假设 OP 当前数据是 Date Amount Item Type 的列类型，就像在
$ cat trans1
#!/usr/bin/env bash
awk '
BEGIN {
OFS = "\t"
}
{
date =
amount =
item =
type =
year = substr(date, 1, 4)
minYear = (NR == 1 || year < minYear ? year : minYear)
maxYear = (NR == 1 || year > maxYear ? year : maxYear)
types_items[type][item]
yearsTypesItems2amounts[year][type][item] +=
years2amounts[year] +=
yearsTypes2amounts[year][type] +=
}
END {
for (year = minYear; year <= maxYear; year++) {
printf "%s%s", OFS, year
}
print ""
print ""
for (type in types_items) {
for (item in types_items[type]) {
printf item, OFS
for (year = minYear; year <= maxYear; year++) {
printf "%s%0.2f", OFS, yearsTypesItems2amounts[year][type][item]
}
print ""
}
printf "sum" type, OFS
for (year = minYear; year <= maxYear; year++) {
printf "%s%0.2f", OFS, yearsTypes2amounts[year][type]
}
print ""
print ""
if (type == "VI") {
printf "sumI", OFS
for (year = minYear; year <= maxYear; year++) {
printf "%s%0.2f", OFS, yearsTypesItems2amounts[year][type][item] # <--- NOTE: item unset here
}
} else if (type == "VE") {
printf "sumE", OFS
for (year = minYear; year <= maxYear; year++) {
printf "%s%0.2f", OFS, yearsTypesItems2amounts[year][type][item] # <--- NOTE: item unset here
}
}
print ORS
}
printf "net"
for (year = minYear; year <= maxYear; year++) {
printf "%s%0.2f", OFS, years2amounts[year]
}
print ""
}
' "${@:--}"
请参阅上面的“注意”，通过重命名练习可以明显看出某些错误所在（看起来您使用了错误的数组，应该使用 yearsTypes2amounts[year][type] 而不是 yearsTypesItems2amounts[year][type][item]).
继续对数组的数组使用 GNU awk，并假设 OP 当前数据是 Date Amount Item Type 的列类型，就像在
$ cat trans1
#!/usr/bin/env bash
awk '
BEGIN {
OFS = "\t"
}
{
date =
amount =
item =
type =
year = substr(date, 1, 4)
minYear = (NR == 1 || year < minYear ? year : minYear)
maxYear = (NR == 1 || year > maxYear ? year : maxYear)
cat = substr(type,2)
ctiys2amounts[cat][type][item][year] += amount
}
END {
for (year = minYear; year <= maxYear; year++) {
printf "%s%s", OFS, year
}
print ORS
for (cat in ctiys2amounts) {
delete catSum
for (type in ctiys2amounts[cat]) {
delete typeSum
for (item in ctiys2amounts[cat][type]) {
printf "%s", item
for (year = minYear; year <= maxYear; year++) {
amount = ctiys2amounts[cat][type][item][year]
printf "%s%0.2f", OFS, amount
typeSum[year] += amount
}
print ""
}
printf "sum%s", type
for (year = minYear; year <= maxYear; year++) {
printf "%s%0.2f", OFS, typeSum[year]
catSum[year] += typeSum[year]
}
print ORS
}
printf "sum%s", cat
for (year = minYear; year <= maxYear; year++) {
printf "%s%0.2f", OFS, catSum[year]
yearSum[year] += catSum[year]
}
print ORS
}
printf "net"
for (year = minYear; year <= maxYear; year++) {
printf "%s%0.2f", OFS, yearSum[year]
}
print ""
}
' "${@:--}"
$ ./trans1 st_in.txt
2015 2016 2017
B 0.00 -2.00 -1.00
D 0.00 -1.00 0.00
sumFE 0.00 -3.00 -1.00
B 0.00 -3.00 0.00
C -4.00 0.00 0.00
D 0.00 -5.00 0.00
sumVE -4.00 -8.00 0.00
sumE -4.00 -11.00 -1.00
A 8.00 0.00 0.00
C 0.00 0.00 8.00
sumFI 8.00 0.00 8.00
A 0.00 0.00 5.00
B 4.00 0.00 0.00
sumVI 4.00 0.00 5.00
sumI 12.00 0.00 13.00
net 8.00 -11.00 12.00