从版本2.25开始，http://yourdomain.com/filter/2017--ABC--Z-400-TT是固定的，因此使用空字节而不是换行来终止输出行。 这对于捕获和处理多行grep匹配很简单（参见示例）
不幸的是，我在生产时遇到了grep版本2.20。 这意味着对于处理\ n终止的日志文件，您无法区分每个输出行的grep-match。
因此我的问题是：
处理每个结果的最有效方法是什么 - 当你遇到2.25之前的版本时，不是使用grep（oz）行命令？
（注意：这是一个更复杂的脚本的一个小例子，需要根据请求处理超过10k的大型日志文件，因此我寻求最有效的＃34;解决方案）
一个简单的例子：
test.log中
flag test1
flag test2
flag test3
test4
test5
flag test6
test7
flag test8
test.sh
#!/bin/bash
#regex explained:
#(?s)enable multiline pattern search
#(flag) capturegroup with pattern indicating new entry
#[[:blank:]] followed by a space
#(.*?) capturegroup for the rest of the entry, non-greedy
#(?=(?:\r\n|[\r\n])(flag)|\z) positive lookahead:
# - stop when the next newline begins with flag
# - OR if last entry is a match: proceed 'till end of entry
regex_multiline="(?s)(flag)[[:blank:]](.*?)(?=(?:\r\n|[\r\n])(flag)|\z)"
logfile="./test.log"
test1(){
#this works only with grep 2.25 or higher,
#which returns a NULL-byte delimiter after each capture
echo start
while IFS= read -r -d '' line ; do
printf '<test>%s</test>\n' "$line"
done < <(grep -Pzo $regex_multiline $logfile)
echo end
}
test2(){
#I need this to work for each match, instead of each line
echo start
while IFS= read -r line ; do
printf '<test>%s</test>\n' "$line"
done < <(grep -Pzo $regex_multiline $logfile)
echo end
}
测试1导致我想要的东西：
start
<test>flag test1</test>
<test>flag test2</test>
<test>flag test3
test4
test5</test>
<test>flag test6
test7
</test>
<test>flag test8</test>
end
测试2结果
start
<test>flag test1</test>
<test>flag test2</test>
<test>flag test3</test>
<test> test4</test>
<test> test5</test>
<test>flag test6</test>
<test></test>
<test>test7</test>
<test> </test>
<test>flag test8</test>
end
答案 0 :(得分：1)
我认为您最好使用#include <iostream>
#include <cstdlib>
#include <ctime>
#include <string>
#include <windows.h>
using namespace std;
//This function is only here to make retrieving random numbers more pretty.
int getrand(int min, int max);
//This function will return what level of rarity the item is as a string.
string getrare();
//This function will return the certification of the item as a string.
string getcert();
//This function will return the crate being opened to a string, and will error check.
string cratePrompt();
//This function will return the specific item received from a specific crate.
string getitem(string crateNum);
int rareCount=0;
int veryrareCount=0;
int importCount=0;
int exoticCount=0;
int blackmarketCount=0;
int certCount=0;
int main(){
unsigned seed;
//This group of text here is to get a seed for randomness.
seed=time(0);
srand(seed);
//This is where the magic happens.
int count=0;
int crateQuantity;
string certTest;
string crateChosen;
crateChosen=(cratePrompt());
cout<<"\nHow many crates do you want to open?\n\n";
cin>>crateQuantity;
cout<<endl;
string itemActual;
while(count<crateQuantity){
certTest=getcert();
if (certTest!="Null"){
cout<<certTest<<" ";
}
cout<<getitem(crateChosen);
cout<<endl;
count++;
}
cout<<"\n\n\nYou got "<<rareCount<<" Rare items!\n";
cout<<"You got "<<veryrareCount<<" Very Rare items!\n";
cout<<"You got "<<importCount<<" Import items!\n";
cout<<"You got "<<exoticCount<<" Exotic items!\n";
cout<<"You got "<<blackmarketCount<<" Black Market items!\n\n";
cout<<certCount<<" of the items were certified!";
}
int getrand(int min, int max){
return ((rand()%(max-min+1))+min);
}
string getrare(){
int rando=(getrand(1,10000));
if ((rando>=1)&&(rando<=5472)){
rareCount++;
return "Rare";
}
else if ((rando>=5473)&&(rando<=8264)){
veryrareCount++;
return "Very Rare";
}
else if ((rando>=8265)&&(rando<=9476)){
importCount++;
return "Import";
}
else if ((rando>=9477)&&(rando<=9868)){
exoticCount++;
return "Exotic";
}
else if ((rando>=9869)&&(rando<=10000)){
blackmarketCount++;
return "Black Market";
}
else{
return "An error has happened";
}
}
string getcert(){
int iscert=getrand(1,100);
int whichcert=getrand(1,15);
if ((iscert>=16)&&(iscert<=100)){
return "Null";
}
else if ((iscert>=1)&&(iscert<=15)){
certCount++;
if (whichcert==1){
return "[Acrobat]";
}
if (whichcert==2){
return "[Aviator]";
}
if (whichcert==3){
return "[Goalkeeper]";
}
if (whichcert==4){
return "[Guardian]";
}
if (whichcert==5){
return "[Juggler]";
}
if (whichcert==6){
return "[Paragon]";
}
if (whichcert==7){
return "[Playmaker]";
}
if (whichcert==8){
return "[Scorer]";
}
if (whichcert==9){
return "[Show-Off]";
}
if (whichcert==10){
return "[Sniper]";
}
if (whichcert==11){
return "[Striker]";
}
if (whichcert==12){
return "[Sweeper]";
}
if (whichcert==13){
return "[Tactician]";
}
if (whichcert==14){
return "[Turtle]";
}
if (whichcert==15){
return "[Victor]";
}
}
else{
return "An error has happened";
}
}
string cratePrompt(){
string answer;
bool valid=0;
while (valid==0){
cout<<"Which crate do you want to open?\n\n";
cin>>answer;
if ((answer=="1")||(answer=="c1")||(answer=="cc1")||(answer=="C1")||(answer=="CC1")){
return "CC1";
}
else if ((answer=="2")||(answer=="c2")||(answer=="cc2")||(answer=="C2")||(answer=="CC2")){
return "CC2";
}
else if ((answer=="3")||(answer=="c3")||(answer=="cc3")||(answer=="C3")||(answer=="CC3")){
return "CC3";
}
else if ((answer=="4")||(answer=="c4")||(answer=="cc4")||(answer=="C4")||(answer=="CC4")){
return "CC4";
}
else{
cout<<"Please enter a valid option (1, 2, 3, or 4)\n";
}
}
}
string getitem(string crateNum){
string rarity = getrare();
//(Rarity)Select is used to select exactly which item they will get out of the possibilities.
int BMSelect= getrand(1,6);
int ESelect= getrand(1,2);
int ISelect= getrand(1,3);
int VRSelect= getrand(1,4);
int RSelect= getrand(1,5);
if (crateNum=="CC1"){
if (rarity=="Rare"){
if(RSelect==1){
return "(Takumi) Combo";
}
else if(RSelect==2){
return "(Breakout) Vice";
}
else if(RSelect==3){
return "(Dominus) Pollo Caliente";
}
else if(RSelect==4){
return "(Dominus) Arcana";
}
else if(RSelect==5){
return "(Breakout) Shibuya";
}
}
else if (rarity=="Very Rare"){
if (VRSelect==1){
return "(Takumi) Anubis";
}
else if (VRSelect==2){
return "(Breakout) Dot Matrix";
}
else if (VRSelect==3){
return "(Dominus) Snakeskin";
}
else if (VRSelect==4){
return "Chakrams";
}
}
else if (rarity=="Import"){
if (ISelect==1){
return "Dominus GT";
}
else if (ISelect==2){
return "Trinity";
}
else if (ISelect==3){
return "Takumi RX-T";
}
}
else if (rarity=="Exotic"){
if (ESelect==1){
return "Photons";
}
else if (ESelect==1){
return "Loopers";
}
}
else if (rarity=="Black Market"){
if (BMSelect==1){
return "Biomass";
}
else if (BMSelect==2){
return "Heatwave";
}
else if (BMSelect==3){
return "Hexed";
}
else if (BMSelect==4){
return "Slipstream";
}
else if (BMSelect==5){
return "Parallax";
}
else if (BMSelect==6){
return "Labyrinth";
}
}
else{
cout<<"AN ERROR HAS OCCURED";
}
}
if (crateNum=="CC2"){
if (rarity=="Rare"){
if(RSelect==1){
return "(Octane) Dragon Lord";
}
if(RSelect==2){
return "(Venom) Nine Lives";
}
if(RSelect==3){
return "(Road Hog) Carbonated";
}
if(RSelect==4){
return "(Takumi) Whizzle";
}
if(RSelect==5){
return "(Merc) Narwhal";
}
}
else if (rarity=="Very Rare"){
if (VRSelect==1){
return "(Octane) Distortion";
}
if (VRSelect==2){
return "(Merc) Warlock";
}
if (VRSelect==3){
return "Polygonal";
}
if (VRSelect==4){
return "(X-Devil) Snakeskin";
}
}
else if (rarity=="Import"){
if (ISelect==1){
return "X-Devil Mk2";
}
if (ISelect==2){
return "Road Hog XL";
}
if (ISelect==3){
return "Pixel Fire";
}
}
else if (rarity=="Exotic"){
if (ESelect==1){
return "Lightnings";
}
if (ESelect==1){
return "Lobos";
}
}
else if (rarity=="Black Market"){
if (BMSelect==1){
return "Biomass";
}
if (BMSelect==2){
return "Heatwave";
}
if (BMSelect==3){
return "Hexed";
}
if (BMSelect==4){
return "Slipstream";
}
if (BMSelect==5){
return "Parallax";
}
if (BMSelect==6){
return "Labyrinth";
}
}
else{
cout<<"AN ERROR HAS OCCURED";
}
}
if (crateNum=="CC3"){
if (rarity=="Rare"){
if(RSelect==1){
return "(Breakout) Falchion";
}
if(RSelect==2){
return "(Breakout) Turbo";
}
if(RSelect==3){
return "(Dominus) Mondo";
}
if(RSelect==4){
return "(Octane) Shisa";
}
if(RSelect==5){
return "(Masamune) Oni";
}
}
else if (rarity=="Very Rare"){
if (VRSelect==1){
return "(Takumi) Distortion";
}
if (VRSelect==2){
return "(Breakout) Snakeskin";
}
if (VRSelect==3){
return "Troikas";
}
if (VRSelect==4){
return "(Octane) MG-88";
}
}
else if (rarity=="Import"){
if (ISelect==1){
return "Breakout Type-S";
}
if (ISelect==2){
return "Hypernova";
}
if (ISelect==3){
return "Dark Matter";
}
}
else if (rarity=="Exotic"){
if (ESelect==1){
return "Pulsus";
}
if (ESelect==1){
return "Discotheques";
}
}
else if (rarity=="Black Market"){
if (BMSelect==1){
return "Biomass";
}
if (BMSelect==2){
return "Heatwave";
}
if (BMSelect==3){
return "Hexed";
}
if (BMSelect==4){
return "Slipstream";
}
if (BMSelect==5){
return "Parallax";
}
if (BMSelect==6){
return "Labyrinth";
}
}
else{
cout<<"AN ERROR HAS OCCURED";
}
}
if (crateNum=="CC4"){
if (rarity=="Rare"){
if(RSelect==1){
return "(Octane) Dragon Lord";
}
if(RSelect==2){
return "(Venom) Nine Lives";
}
if(RSelect==3){
return "(Road Hog) Carbonated";
}
if(RSelect==4){
return "(Takumi) Whizzle";
}
if(RSelect==5){
return "(Merc) Narwhal";
}
}
else if (rarity=="Very Rare"){
if (VRSelect==1){
return "(Octane) Distortion";
}
if (VRSelect==2){
return "(Merc) Warlock";
}
if (VRSelect==3){
return "Polygonal";
}
if (VRSelect==4){
return "(X-Devil) Snakeskin";
}
}
else if (rarity=="Import"){
if (ISelect==1){
return "X-Devil Mk2";
}
if (ISelect==2){
return "Road Hog XL";
}
if (ISelect==3){
return "Pixel Fire";
}
}
else if (rarity=="Exotic"){
if (ESelect==1){
return "Lightnings";
}
if (ESelect==1){
return "Lobos";
}
}
else if (rarity=="Black Market"){
if (BMSelect==1){
return "Biomass";
}
if (BMSelect==2){
return "Heatwave";
}
if (BMSelect==3){
return "Hexed";
}
if (BMSelect==4){
return "Slipstream";
}
if (BMSelect==5){
return "Parallax";
}
if (BMSelect==6){
return "Labyrinth";
}
}
else{
cout<<"AN ERROR HAS OCCURED";
}
}
}
代替perl。您可以使用几乎未经修改的 1 正则表达式，只需将其替换为grep 2 ：
\1\x00 1 你的正则表达式有点奇怪，捕获组在grep命令的上下文中没有做任何事情（比如regex_multiline="(?s)(flag[[:blank:]].*?)(?=(?:\r\n|[\r\n])flag|\z)"
perl -0777 -pe "s/$regex_multiline/\1\x00/g" < "$logfile"
）。我只是将您想要匹配的整个部分放入一个组中，以便它与替换部分中的(flag)相对应。如果需要调整/我错过了什么。
2 使用\1（对于＆＃34;匹配组1＆＃34;，＆＃34;空字节＆＃34;）实际上也有效但看起来很亲切令人困惑的。
答案 1 :(得分：0)
我找到了解决方案。我想，这有点像黑客攻击，但它与grep版本2.20及更高版本一致。虽然不要使用它与grep 2.25及以上。 这是grep与参数-zon的组合： - z（将输入视为一组行，每个行以零字节结尾） - o（仅打印匹配行的匹配（非空）部分） - n（在输入文件中使用从1开始的行号前缀每行输出。）
此组合将在每次新比赛开始时输出“1：”。总是。 （不确定它是grep中的错误，还是设计错误，但选项-z和-o确实有意义）
1:flag test1
1:flag test2
1:flag test3
test4
test5
1:flag test6
test7
1:flag test8
因此，知道这一点，这将导致以下的find-and-replace函数将替换以1开头的每一行：使用空字节字符。请注意，每行末尾都需要一个空字节字符，因此我们必须手动为最后一行添加一个字符！
这可以通过以下方式完成：
sed -e's / ^ 1：/ \ x0 / g'| sed -e'$ a \ x0' 要么 awk'{gsub（/ ^ 1：/，“\ x0”）;} 1'| sed -e'$ a \ x0'
（我认为sed对于这种操作更有效/更快，但不要因此而限制我。）
test2(){
#This finally works!
echo start
while IFS= read -r -d '' line ; do
printf '<test>%s</test>\n' "$line"
done < <(grep -Pzon $regex_multiline $logfile | sed -e 's/^1:/\x0/g' | sed -e '$a\\x0' )
echo end
}