对文件中数据块内的行进行排序。根据列中的索引将一个块中的行移动到数据块中的不同位置
Sort rows within data blocks in a file. Move rows in one block to a different location in the data block based on index in a column
我的文件中有如下排列的数据。这只显示了两个数据blocks/iterations.
21 ! <-- This is the number of lines of data in the data block/iteration. It never changes.
Linkages. Iteration:1_1010 ! <-- This number does not always increase by 5 like in this example, but always increases.
A 1.010 -3.582 -3.135
B 0.730 -4.428 -3.854
B -3.883 4.671 0.010
A -0.223 2.522 -4.893
B 2.769 4.634 0.179
B -2.024 -3.640 -1.032
A 4.613 3.914 1.567
B 2.746 -0.545 1.430
B -0.532 3.380 -2.107
C 3.944 2.513 -5.172
C -4.669 1.056 2.747
C 0.645 0.001 -3.737
C -2.875 -1.233 -0.538
C 4.279 -5.187 -2.820
C 1.067 -2.279 2.021
C 2.667 -1.558 0.588
C 3.628 -0.025 2.464
C -0.023 1.717 1.175
C 0.925 -1.548 2.273
C 1.152 2.914 1.039
C 0.878 -0.445 -0.948
21
Linkages. Iteration:1_1015
A 1.010 -3.582 -3.135
B 0.730 -4.428 -3.854
B -3.883 4.671 0.010
A -0.223 2.522 -4.893
B 2.769 4.634 0.179
B -2.024 -3.640 -1.032
A 4.613 3.914 1.567
B 2.746 -0.545 1.430
B -0.532 3.380 -2.107
C 3.944 2.513 -5.172
C -4.669 1.056 2.747
C 0.645 0.001 -3.737
C -2.875 -1.233 -0.538
C 4.279 -5.187 -2.820
C 1.067 -2.279 2.021
C 2.667 -1.558 0.588
C 3.628 -0.025 2.464
C -0.023 1.717 1.175
C 0.925 -1.548 2.273
C 1.152 2.914 1.039
C 0.878 -0.445 -0.948
我需要做的是重新分配“C”行。具体来说，我需要将“C”行分成四个块，然后将第一组 C 行移动到第一组“ABB”行下方。这是一个数据 block/iteration 的示例（我想对文件中的所有数据 blocks/iterations 做完全相同的事情）：
21
Linkages. Iteration:1_1010
A 1.010 -3.582 -3.135
B 0.730 -4.428 -3.854
B -3.883 4.671 0.010
C 3.944 2.513 -5.172
C -4.669 1.056 2.747
C 0.645 0.001 -3.737
C -2.875 -1.233 -0.538
A -0.223 2.522 -4.893
B 2.769 4.634 0.179
B -2.024 -3.640 -1.032
C 4.279 -5.187 -2.820
C 1.067 -2.279 2.021
C 2.667 -1.558 0.588
C 3.628 -0.025 2.464
A 4.613 3.914 1.567
B 2.746 -0.545 1.430
B -0.532 3.380 -2.107
C -0.023 1.717 1.175
C 0.925 -1.548 2.273
C 1.152 2.914 1.039
C 0.878 -0.445 -0.948
我一直在 bash 尝试使用“排序”来做到这一点，但没有取得太大进展。我发现按列索引（如我的第一列）排序的一般方法是这样做：
sort -n -k1 file
我还发现了这个 post (https://unix.stackexchange.com/questions/99582/sorting-blocks-of-lines)，其中第二个答案使用“拆分”将文件拆分为由四行组成的块：
split -a 6 -l 4 input_file my_prefix_
但我不知道如何用数据 block/iteration 移动四行。如果有人知道解释这一点的资源，那将是很好的发现。
这是 GNU awk 中的一个。只测试了你的数据，我不知道如果数据不完美会发生什么，可能你会得到空行。此外，它使用 getline，它可能有问题，我没有检查它的 return 值，也没有检查任何其他东西——但从好的方面来说，这是一个好的开始开始练习你的 awk 技能...:D
$ gawk '
BEGIN {
o="A,B,B,C,C,C,C" # the predefined order
n=split(o,p,/,/) # split to p array to be fetched
}
{
c=d= # count of input lines
while(c-->=0) { # keep reading til the reading is done
getline
a[][++i[]]=[=10=] # hash records to a 2-d array
}
print d
print a["Linkages."][1] # this may help in understanding the a array
do { # once required amount is hashed
for(j=1;j<=n;j++) { # use the BEGIN defined index order
print a[p[j]][++k[p[j]]] # and output
}
} while((d-=n)>0)
delete a;delete i;delete k # regroup for next batch
}' file
此解决方案的工作原理是将每个 21 行输入数据块分解为一个二维数组，每个子维度由 7 行组成 (A,B,B,C,C,C,C)：
blocks[1][ 1] = A record
blocks[1][2-3] = B records
blocks[1][4-7] = C records
blocks[2][ 1] = A record
blocks[2][2-3] = B records
blocks[2][4-7] = C records
blocks[3][ 1] = A record
blocks[3][2-3] = B records
blocks[3][4-7] = C records
一个awk解决方案使用了这个二维数组的想法；我们将在处理 A 和 B 记录时填充数组，然后用 C 记录回填数组（即填充空白）：
awk '
# function to print current array contents to stdout, then reset data structures for next block of lines
function print_blocks() {
for (i=1; i<=a; i++) # loop through first dimension indices
for (j=1; j<=7; j++) # loop through second dimension indices
print blocks[i][j]
delete blocks # clear array
a=0 # reset first dimension index
cblock=0 # reset C block processing flag
}
NF == 1 { # single field on line, eg, "21" ?
print_blocks() # flush previously populated array
print # print current line
next # skip to next line
}
/Linkages/ { print # print current line
next # skip to next line
}
== "A" { # "A" record?
blocks[++a][1]=[=11=] # store current line in array
# and reset second dimension indexes ...
b=2 # for B records
c=4 # for C records
next
}
== "B" { # "B" record?
blocks[a][b++]=[=11=] # store current line in array
next
}
== "C" { # "C" record?
if (cblock==0) # if first C record then:
{ a=0 # reset first dimension index
cblock=1 # set flag to skip this logic for rest of C records
}
if (c==4) a++ # for each new set of 4x C records increment first dimension index
blocks[a][c]=[=11=] # store current line in array
c++ # increment second dimension index but ...
if (c>7) c=4 # make sure second dimension index is always in the range 4-7
next
}
END { print_blocks() } # flush the last set of array data to stdout
' data.txt
注释：
- 取决于匹配样本数据的输入数据（即
A,B,B,A,B,B,A,B,B,C,C,....）；如果输入顺序不同，则此代码可能不会生成所需的输出
- 对于给定的样本，这显然是硬编码的（即 21 行输入数据，
A,B,B,C,C,C,C 的每个输出块 7 行）
- 可以修改代码以处理一组更动态的输入数据（但不是所有代码都如此吗？）
- 可以删除注释以整理代码
- 需要
GNU awk 来实现二维数组（又名数组）
针对给定的示例数据 (data.txt)，以上生成：
21
Linkages. Iteration:1_1010
A 1.010 -3.582 -3.135
B 0.730 -4.428 -3.854
B -3.883 4.671 0.010
C 3.944 2.513 -5.172
C -4.669 1.056 2.747
C 0.645 0.001 -3.737
C -2.875 -1.233 -0.538
A -0.223 2.522 -4.893
B 2.769 4.634 0.179
B -2.024 -3.640 -1.032
C 4.279 -5.187 -2.820
C 1.067 -2.279 2.021
C 2.667 -1.558 0.588
C 3.628 -0.025 2.464
A 4.613 3.914 1.567
B 2.746 -0.545 1.430
B -0.532 3.380 -2.107
C -0.023 1.717 1.175
C 0.925 -1.548 2.273
C 1.152 2.914 1.039
C 0.878 -0.445 -0.948
21
Linkages. Iteration:1_1015
A 1.010 -3.582 -3.135
B 0.730 -4.428 -3.854
B -3.883 4.671 0.010
C 3.944 2.513 -5.172
C -4.669 1.056 2.747
C 0.645 0.001 -3.737
C -2.875 -1.233 -0.538
A -0.223 2.522 -4.893
B 2.769 4.634 0.179
B -2.024 -3.640 -1.032
C 4.279 -5.187 -2.820
C 1.067 -2.279 2.021
C 2.667 -1.558 0.588
C 3.628 -0.025 2.464
A 4.613 3.914 1.567
B 2.746 -0.545 1.430
B -0.532 3.380 -2.107
C -0.023 1.717 1.175
C 0.925 -1.548 2.273
C 1.152 2.914 1.039
C 0.878 -0.445 -0.948
在每个 Unix 机器上的任何 shell 中使用任何 awk：
$ cat tst.awk
~ /^[ABC]$/ {
vals[++numVals] = [=10=]
next
}
{
prtVals()
print
}
END { prtVals() }
function prtVals( row,valNr,blocks,numBlocks,blockNr,numCs) {
if ( numVals != 0 ) {
for (valNr=1; valNr<=numVals; valNr++) {
row = vals[valNr]
split(row,f)
if ( f[1] == "A" ) {
++numBlocks
}
if ( f[1] == "C" ) {
if ( (++numCs % 4) == 1 ) {
blockNr++
}
blocks[blockNr] = blocks[blockNr] row ORS
}
else {
blocks[numBlocks] = blocks[numBlocks] row ORS
}
}
for (blockNr=1; blockNr<=numBlocks; blockNr++) {
printf "%s", blocks[blockNr]
}
delete vals
numVals = 0
}
}
$ awk -f tst.awk file
21 ! <-- This is the number of lines of data in the data block/iteration. It never changes.
Linkages. Iteration:1_1010 ! <-- This number does not always increase by 5 like in this example, but always increases.
A 1.010 -3.582 -3.135
B 0.730 -4.428 -3.854
B -3.883 4.671 0.010
C 3.944 2.513 -5.172
C -4.669 1.056 2.747
C 0.645 0.001 -3.737
C -2.875 -1.233 -0.538
A -0.223 2.522 -4.893
B 2.769 4.634 0.179
B -2.024 -3.640 -1.032
C 4.279 -5.187 -2.820
C 1.067 -2.279 2.021
C 2.667 -1.558 0.588
C 3.628 -0.025 2.464
A 4.613 3.914 1.567
B 2.746 -0.545 1.430
B -0.532 3.380 -2.107
C -0.023 1.717 1.175
C 0.925 -1.548 2.273
C 1.152 2.914 1.039
C 0.878 -0.445 -0.948
21
Linkages. Iteration:1_1015
A 1.010 -3.582 -3.135
B 0.730 -4.428 -3.854
B -3.883 4.671 0.010
C 3.944 2.513 -5.172
C -4.669 1.056 2.747
C 0.645 0.001 -3.737
C -2.875 -1.233 -0.538
A 1.010 -3.582 -3.135
B 0.730 -4.428 -3.854
B -3.883 4.671 0.010
C 3.944 2.513 -5.172
C -4.669 1.056 2.747
C 0.645 0.001 -3.737
C -2.875 -1.233 -0.538
A -0.223 2.522 -4.893
B 2.769 4.634 0.179
B -2.024 -3.640 -1.032
C 4.279 -5.187 -2.820
C 1.067 -2.279 2.021
C 2.667 -1.558 0.588
C 3.628 -0.025 2.464
A -0.223 2.522 -4.893
B 2.769 4.634 0.179
B -2.024 -3.640 -1.032
C 4.279 -5.187 -2.820
C 1.067 -2.279 2.021
C 2.667 -1.558 0.588
C 3.628 -0.025 2.464
A 4.613 3.914 1.567
B 2.746 -0.545 1.430
B -0.532 3.380 -2.107
C -0.023 1.717 1.175
C 0.925 -1.548 2.273
C 1.152 2.914 1.039
C 0.878 -0.445 -0.948
A 4.613 3.914 1.567
B 2.746 -0.545 1.430
B -0.532 3.380 -2.107
C -0.023 1.717 1.175
C 0.925 -1.548 2.273
C 1.152 2.914 1.039
C 0.878 -0.445 -0.948
我的文件中有如下排列的数据。这只显示了两个数据blocks/iterations.
21 ! <-- This is the number of lines of data in the data block/iteration. It never changes.
Linkages. Iteration:1_1010 ! <-- This number does not always increase by 5 like in this example, but always increases.
A 1.010 -3.582 -3.135
B 0.730 -4.428 -3.854
B -3.883 4.671 0.010
A -0.223 2.522 -4.893
B 2.769 4.634 0.179
B -2.024 -3.640 -1.032
A 4.613 3.914 1.567
B 2.746 -0.545 1.430
B -0.532 3.380 -2.107
C 3.944 2.513 -5.172
C -4.669 1.056 2.747
C 0.645 0.001 -3.737
C -2.875 -1.233 -0.538
C 4.279 -5.187 -2.820
C 1.067 -2.279 2.021
C 2.667 -1.558 0.588
C 3.628 -0.025 2.464
C -0.023 1.717 1.175
C 0.925 -1.548 2.273
C 1.152 2.914 1.039
C 0.878 -0.445 -0.948
21
Linkages. Iteration:1_1015
A 1.010 -3.582 -3.135
B 0.730 -4.428 -3.854
B -3.883 4.671 0.010
A -0.223 2.522 -4.893
B 2.769 4.634 0.179
B -2.024 -3.640 -1.032
A 4.613 3.914 1.567
B 2.746 -0.545 1.430
B -0.532 3.380 -2.107
C 3.944 2.513 -5.172
C -4.669 1.056 2.747
C 0.645 0.001 -3.737
C -2.875 -1.233 -0.538
C 4.279 -5.187 -2.820
C 1.067 -2.279 2.021
C 2.667 -1.558 0.588
C 3.628 -0.025 2.464
C -0.023 1.717 1.175
C 0.925 -1.548 2.273
C 1.152 2.914 1.039
C 0.878 -0.445 -0.948
我需要做的是重新分配“C”行。具体来说，我需要将“C”行分成四个块，然后将第一组 C 行移动到第一组“ABB”行下方。这是一个数据 block/iteration 的示例（我想对文件中的所有数据 blocks/iterations 做完全相同的事情）：
21
Linkages. Iteration:1_1010
A 1.010 -3.582 -3.135
B 0.730 -4.428 -3.854
B -3.883 4.671 0.010
C 3.944 2.513 -5.172
C -4.669 1.056 2.747
C 0.645 0.001 -3.737
C -2.875 -1.233 -0.538
A -0.223 2.522 -4.893
B 2.769 4.634 0.179
B -2.024 -3.640 -1.032
C 4.279 -5.187 -2.820
C 1.067 -2.279 2.021
C 2.667 -1.558 0.588
C 3.628 -0.025 2.464
A 4.613 3.914 1.567
B 2.746 -0.545 1.430
B -0.532 3.380 -2.107
C -0.023 1.717 1.175
C 0.925 -1.548 2.273
C 1.152 2.914 1.039
C 0.878 -0.445 -0.948
我一直在 bash 尝试使用“排序”来做到这一点，但没有取得太大进展。我发现按列索引（如我的第一列）排序的一般方法是这样做：
sort -n -k1 file
我还发现了这个 post (https://unix.stackexchange.com/questions/99582/sorting-blocks-of-lines)，其中第二个答案使用“拆分”将文件拆分为由四行组成的块：
split -a 6 -l 4 input_file my_prefix_
但我不知道如何用数据 block/iteration 移动四行。如果有人知道解释这一点的资源，那将是很好的发现。
这是 GNU awk 中的一个。只测试了你的数据，我不知道如果数据不完美会发生什么，可能你会得到空行。此外，它使用 getline，它可能有问题，我没有检查它的 return 值，也没有检查任何其他东西——但从好的方面来说，这是一个好的开始开始练习你的 awk 技能...:D
$ gawk '
BEGIN {
o="A,B,B,C,C,C,C" # the predefined order
n=split(o,p,/,/) # split to p array to be fetched
}
{
c=d= # count of input lines
while(c-->=0) { # keep reading til the reading is done
getline
a[][++i[]]=[=10=] # hash records to a 2-d array
}
print d
print a["Linkages."][1] # this may help in understanding the a array
do { # once required amount is hashed
for(j=1;j<=n;j++) { # use the BEGIN defined index order
print a[p[j]][++k[p[j]]] # and output
}
} while((d-=n)>0)
delete a;delete i;delete k # regroup for next batch
}' file
此解决方案的工作原理是将每个 21 行输入数据块分解为一个二维数组，每个子维度由 7 行组成 (A,B,B,C,C,C,C)：
blocks[1][ 1] = A record
blocks[1][2-3] = B records
blocks[1][4-7] = C records
blocks[2][ 1] = A record
blocks[2][2-3] = B records
blocks[2][4-7] = C records
blocks[3][ 1] = A record
blocks[3][2-3] = B records
blocks[3][4-7] = C records
一个awk解决方案使用了这个二维数组的想法；我们将在处理 A 和 B 记录时填充数组，然后用 C 记录回填数组（即填充空白）：
awk '
# function to print current array contents to stdout, then reset data structures for next block of lines
function print_blocks() {
for (i=1; i<=a; i++) # loop through first dimension indices
for (j=1; j<=7; j++) # loop through second dimension indices
print blocks[i][j]
delete blocks # clear array
a=0 # reset first dimension index
cblock=0 # reset C block processing flag
}
NF == 1 { # single field on line, eg, "21" ?
print_blocks() # flush previously populated array
print # print current line
next # skip to next line
}
/Linkages/ { print # print current line
next # skip to next line
}
== "A" { # "A" record?
blocks[++a][1]=[=11=] # store current line in array
# and reset second dimension indexes ...
b=2 # for B records
c=4 # for C records
next
}
== "B" { # "B" record?
blocks[a][b++]=[=11=] # store current line in array
next
}
== "C" { # "C" record?
if (cblock==0) # if first C record then:
{ a=0 # reset first dimension index
cblock=1 # set flag to skip this logic for rest of C records
}
if (c==4) a++ # for each new set of 4x C records increment first dimension index
blocks[a][c]=[=11=] # store current line in array
c++ # increment second dimension index but ...
if (c>7) c=4 # make sure second dimension index is always in the range 4-7
next
}
END { print_blocks() } # flush the last set of array data to stdout
' data.txt
注释：
- 取决于匹配样本数据的输入数据（即
A,B,B,A,B,B,A,B,B,C,C,....）；如果输入顺序不同，则此代码可能不会生成所需的输出 - 对于给定的样本，这显然是硬编码的（即 21 行输入数据，
A,B,B,C,C,C,C的每个输出块 7 行） - 可以修改代码以处理一组更动态的输入数据（但不是所有代码都如此吗？）
- 可以删除注释以整理代码
- 需要
GNU awk来实现二维数组（又名数组）
针对给定的示例数据 (data.txt)，以上生成：
21
Linkages. Iteration:1_1010
A 1.010 -3.582 -3.135
B 0.730 -4.428 -3.854
B -3.883 4.671 0.010
C 3.944 2.513 -5.172
C -4.669 1.056 2.747
C 0.645 0.001 -3.737
C -2.875 -1.233 -0.538
A -0.223 2.522 -4.893
B 2.769 4.634 0.179
B -2.024 -3.640 -1.032
C 4.279 -5.187 -2.820
C 1.067 -2.279 2.021
C 2.667 -1.558 0.588
C 3.628 -0.025 2.464
A 4.613 3.914 1.567
B 2.746 -0.545 1.430
B -0.532 3.380 -2.107
C -0.023 1.717 1.175
C 0.925 -1.548 2.273
C 1.152 2.914 1.039
C 0.878 -0.445 -0.948
21
Linkages. Iteration:1_1015
A 1.010 -3.582 -3.135
B 0.730 -4.428 -3.854
B -3.883 4.671 0.010
C 3.944 2.513 -5.172
C -4.669 1.056 2.747
C 0.645 0.001 -3.737
C -2.875 -1.233 -0.538
A -0.223 2.522 -4.893
B 2.769 4.634 0.179
B -2.024 -3.640 -1.032
C 4.279 -5.187 -2.820
C 1.067 -2.279 2.021
C 2.667 -1.558 0.588
C 3.628 -0.025 2.464
A 4.613 3.914 1.567
B 2.746 -0.545 1.430
B -0.532 3.380 -2.107
C -0.023 1.717 1.175
C 0.925 -1.548 2.273
C 1.152 2.914 1.039
C 0.878 -0.445 -0.948
在每个 Unix 机器上的任何 shell 中使用任何 awk：
$ cat tst.awk
~ /^[ABC]$/ {
vals[++numVals] = [=10=]
next
}
{
prtVals()
print
}
END { prtVals() }
function prtVals( row,valNr,blocks,numBlocks,blockNr,numCs) {
if ( numVals != 0 ) {
for (valNr=1; valNr<=numVals; valNr++) {
row = vals[valNr]
split(row,f)
if ( f[1] == "A" ) {
++numBlocks
}
if ( f[1] == "C" ) {
if ( (++numCs % 4) == 1 ) {
blockNr++
}
blocks[blockNr] = blocks[blockNr] row ORS
}
else {
blocks[numBlocks] = blocks[numBlocks] row ORS
}
}
for (blockNr=1; blockNr<=numBlocks; blockNr++) {
printf "%s", blocks[blockNr]
}
delete vals
numVals = 0
}
}
$ awk -f tst.awk file
21 ! <-- This is the number of lines of data in the data block/iteration. It never changes.
Linkages. Iteration:1_1010 ! <-- This number does not always increase by 5 like in this example, but always increases.
A 1.010 -3.582 -3.135
B 0.730 -4.428 -3.854
B -3.883 4.671 0.010
C 3.944 2.513 -5.172
C -4.669 1.056 2.747
C 0.645 0.001 -3.737
C -2.875 -1.233 -0.538
A -0.223 2.522 -4.893
B 2.769 4.634 0.179
B -2.024 -3.640 -1.032
C 4.279 -5.187 -2.820
C 1.067 -2.279 2.021
C 2.667 -1.558 0.588
C 3.628 -0.025 2.464
A 4.613 3.914 1.567
B 2.746 -0.545 1.430
B -0.532 3.380 -2.107
C -0.023 1.717 1.175
C 0.925 -1.548 2.273
C 1.152 2.914 1.039
C 0.878 -0.445 -0.948
21
Linkages. Iteration:1_1015
A 1.010 -3.582 -3.135
B 0.730 -4.428 -3.854
B -3.883 4.671 0.010
C 3.944 2.513 -5.172
C -4.669 1.056 2.747
C 0.645 0.001 -3.737
C -2.875 -1.233 -0.538
A 1.010 -3.582 -3.135
B 0.730 -4.428 -3.854
B -3.883 4.671 0.010
C 3.944 2.513 -5.172
C -4.669 1.056 2.747
C 0.645 0.001 -3.737
C -2.875 -1.233 -0.538
A -0.223 2.522 -4.893
B 2.769 4.634 0.179
B -2.024 -3.640 -1.032
C 4.279 -5.187 -2.820
C 1.067 -2.279 2.021
C 2.667 -1.558 0.588
C 3.628 -0.025 2.464
A -0.223 2.522 -4.893
B 2.769 4.634 0.179
B -2.024 -3.640 -1.032
C 4.279 -5.187 -2.820
C 1.067 -2.279 2.021
C 2.667 -1.558 0.588
C 3.628 -0.025 2.464
A 4.613 3.914 1.567
B 2.746 -0.545 1.430
B -0.532 3.380 -2.107
C -0.023 1.717 1.175
C 0.925 -1.548 2.273
C 1.152 2.914 1.039
C 0.878 -0.445 -0.948
A 4.613 3.914 1.567
B 2.746 -0.545 1.430
B -0.532 3.380 -2.107
C -0.023 1.717 1.175
C 0.925 -1.548 2.273
C 1.152 2.914 1.039
C 0.878 -0.445 -0.948