循环不会遍历 bash 中的日志文件
The loop does not iterate over the log file in bash
欢迎大家，
无奈，只好求助
我正在尝试编写一个简单的脚本来显示基于日志文件的过程时间。但是，我的循环无法正常工作。
fun.sh
STARTPROCEDURES="START Search"
ENDPROCEDURES="END Search"
start() {
start="$(grep "$STARTPROCEDURES" s.log | cut -d ' ' -f2)"
hours="$(date -d $start '+%H')"
minutes="$(date -d $start '+%M')"
seconds="$(date -d $start '+%S')"
milliseconds=$(($(date -d $start +%N | sed 's/^0*//')/1000000))
starttime=$(((hours * 360000000 + minutes * 60000 + seconds * 1000 + milliseconds)))
echo "Begins: " $starttime
}
end() {
end="$(grep "$ENDPROCEDURES" s.log | cut -d ' ' -f2)"
hours="$(date -d $end '+%H')"
minutes="$(date -d $end '+%M')"
seconds="$(date -d $end '+%S')"
milliseconds=$(($(date -d $end +%N | sed 's/^0*//')/1000000))
endtime=$(((hours * 360000000 + minutes * 60000 + seconds * 1000 + milliseconds)))
echo "Ends: " $endtime
}
difference() {
echo "DIFFERENCE -----"
difference=$((endtime - starttime))
echo "The difference is" $difference "milliseconds"
}
calculate() {
start
end
difference
}
while IFS= read -r line || [[ -n $line ]]; do
calculate
echo "-------"
done < s.log
s.log:
2019-02-22 06:27:06,857 INFO [ProcedureUtil] - (user1,14) START Search
2019-02-22 06:27:06,939 INFO [ProcedureUtil] - (user1,14) END Search
2019-02-22 07:28:16,088 INFO [ProcedureUtil] - (user1,67) START Search
2019-02-22 07:28:16,121 INFO [ProcedureUtil] - (user1,67) END Search
bash fun.sh
休斯顿，我们有问题了。
控制台输出：
date: extra operand ‘+%H’
Try 'date --help' for more information.
date: extra operand ‘+%M’
Try 'date --help' for more information.
date: extra operand ‘+%S’
Try 'date --help' for more information.
date: extra operand ‘+%N’
Try 'date --help' for more information.
fun.sh: line 21: /1000000: syntax error: operand expected (error token is "/1000000")
应该是这样的
Begins: 2161626857
Ends: 2161626939
DIFFERENCE -----
The difference is 82 milliseconds
------
Begins: 216162343
Ends: 216162355
DIFFERENCE -----
The difference is 162 milliseconds
如果 s.log 是：
2019-02-22 06:27:06,857 INFO [ProcedureUtil] - (user1,14) START Search
2019-02-22 06:27:06,939 INFO [ProcedureUtil] - (user1,14) END Search
控制台输出：
Begins: 2161626857
Ends: 2161626939
DIFFERENCE -----
The difference is 82 milliseconds
-------
Begins: 2161626857
Ends: 2161626939
DIFFERENCE -----
The difference is 82 milliseconds
-------
两次一样，应该是一次。
求助我会很优雅
我正在考虑一个案例：
其中 s.log 是：
2019-02-22 06:27:06,857 INFO [ProcedureUtil] - (user1,14) START Search
2019-02-22 06:27:06,939 INFO [ProcedureUtil] - (user1,14) END Search
2019-02-22 07:28:16,088 INFO [ProcedureUtil] - (user1,67) START Split
2019-02-22 07:28:16,121 INFO [ProcedureUtil] - (user1,67) END Split
2019-02-22 07:28:16,088 INFO [ProcedureUtil] - (user1,67) START Search
2019-02-22 07:28:16,121 INFO [ProcedureUtil] - (user1,67) END Search
2019-02-25 20:59:59,999 INFO [ProcedureUtil] - (user1,17) START Search
2019-02-25 02:59:59,999 INFO [ProcedureUtil] - (user1,18) START Search
错误的解决方案：
COUNTER=0
while IFS= read -r line || [[ -n $line ]]; do
if [[ $line == *"START Search"* ]]; then
start=$(time2millis "$line")
echo "Begins: $start"
elif [[ $line == *"END Search"* ]]; then
end=$(time2millis "$line")
echo "Ends: $end"
# Assume every END has a preceding START
difference "$start" "$end"
else
COUNTER=$((COUNTER+1))
echo "$COUNTER"
fi
done < s.log
控制台输出：
Begins: 23226857
Ends: 23226939
DIFFERENCE -----
The difference is 82 milliseconds
------
Begins: 26896088
Ends: 26896121
DIFFERENCE -----
The difference is 33 milliseconds
------
Begins: 75599999
Begins: 10799999
顾虑：
- 你正在逐行阅读文件，但你没有在任何地方使用
$line：你正在 grep 整个文件
- 因为 grep 调用 returns 多次匹配，你的
$start 变量包含换行符
因为你没有在 date 调用中引用变量，所以看起来像这样
date -d 06:27:06,857 07:28:16,088 '+%H'
显然参数太多了 date
大多数错误都可以通过简化代码来消除。请注意，您的 start 和 end 函数在功能上是相同的。
difference() {
local diff=$(( - ))
printf 'DIFFERENCE -----\nThe difference is %d milliseconds\n------\n' "$diff"
}
time2millis() {
local time=$(echo "" | cut -d ' ' -f 2)
IFS=:, read -r hh mm ss nnn <<<"$time"
# be aware of invalid octal numbers 08 and 09:
# each component of the time must be handled as a decimal number
echo "$(( (((10#$hh) * 60 + 10#$mm) * 60 + 10#$ss) * 1000 + 10#$nnn ))"
}
while IFS= read -r line || [[ -n $line ]]; do
if [[ $line == *"START Search"* ]]; then
start=$(time2millis "$line")
echo "Begins: $start"
elif [[ $line == *"END Search"* ]]; then
end=$(time2millis "$line")
echo "Ends: $end"
# Assume every END has a preceding START
difference "$start" "$end"
fi
done < s.log
进一步发展：
这里不考虑END发生在第二天的情况
2019-02-28 23:59:59,999 blah blah START Search
2019-03-01 00:00:00,001 blah blah END Search
您将看到 -86399998
- 您没有考虑夏令时（除非您的日志以 UTC 格式记录时间戳）。
我认为可以通过使用 date 解析日期和时间来解决这些问题：
time2millis() {
local epoch=$( date -d "$(echo "" | cut -d ' ' -f 1,2)" '+%s' )
local millis=0
[[ =~ ,([0-9]+) ]] && millis=${BASH_REMATCH[1]}
echo "$(( $epoch * 1000 + 10#$millis ))"
}
完整的解决方案：
difference() {
local diff=$(( - ))
while (( diff < 0 )); do (( diff += 86400000 )); done
printf "DIFFERENCE -----\nThe difference is %d milliseconds\n------\n" "$diff"
}
# returns milliseconds since 1970-01-01 00:00:00 UTC
time2millis() {
local epoch=$( date -d "$(echo "" | cut -d ' ' -f 1,2)" '+%s' )
local millis=0
[[ =~ ,([0-9]+) ]] && millis=${BASH_REMATCH[1]}
echo "$(( $epoch * 1000 + 10#$millis ))"
}
declare -A startTime startLine
while IFS= read -r line || [[ -n $line ]]; do
read -a words <<<"$line"
key=${words[5]}
if [[ "${words[6]} ${words[7]}" == "START Search" ]]; then
startLine[$key]=$line
startTime[$key]=$(time2millis "$line")
elif [[ "${words[6]} ${words[7]}" == "END Search" ]]; then
if [[ -z ${startTime[$key]} ]]; then
echo "END seen with no START: $line"
else
end=$(time2millis "$line")
difference "$key" "${startTime[$key]}" "$end"
unset startLine[$key] startTime[$key]
fi
fi
done < s.log
echo "Searches STARTed but not ENDed:"
printf "%s\n" "${startLine[@]}"
给定你的输入数据，输出
DIFFERENCE (user1,14) -----
The difference is 82 milliseconds
------
DIFFERENCE (user1,67) -----
The difference is 33 milliseconds
------
Searches STARTed but not ENDed:
2019-02-25 02:59:59,999 INFO [ProcedureUtil] - (user1,18) START Search
2019-02-25 20:59:59,999 INFO [ProcedureUtil] - (user1,17) START Search
欢迎大家，
无奈，只好求助
我正在尝试编写一个简单的脚本来显示基于日志文件的过程时间。但是，我的循环无法正常工作。
fun.sh
STARTPROCEDURES="START Search"
ENDPROCEDURES="END Search"
start() {
start="$(grep "$STARTPROCEDURES" s.log | cut -d ' ' -f2)"
hours="$(date -d $start '+%H')"
minutes="$(date -d $start '+%M')"
seconds="$(date -d $start '+%S')"
milliseconds=$(($(date -d $start +%N | sed 's/^0*//')/1000000))
starttime=$(((hours * 360000000 + minutes * 60000 + seconds * 1000 + milliseconds)))
echo "Begins: " $starttime
}
end() {
end="$(grep "$ENDPROCEDURES" s.log | cut -d ' ' -f2)"
hours="$(date -d $end '+%H')"
minutes="$(date -d $end '+%M')"
seconds="$(date -d $end '+%S')"
milliseconds=$(($(date -d $end +%N | sed 's/^0*//')/1000000))
endtime=$(((hours * 360000000 + minutes * 60000 + seconds * 1000 + milliseconds)))
echo "Ends: " $endtime
}
difference() {
echo "DIFFERENCE -----"
difference=$((endtime - starttime))
echo "The difference is" $difference "milliseconds"
}
calculate() {
start
end
difference
}
while IFS= read -r line || [[ -n $line ]]; do
calculate
echo "-------"
done < s.log
s.log:
2019-02-22 06:27:06,857 INFO [ProcedureUtil] - (user1,14) START Search
2019-02-22 06:27:06,939 INFO [ProcedureUtil] - (user1,14) END Search
2019-02-22 07:28:16,088 INFO [ProcedureUtil] - (user1,67) START Search
2019-02-22 07:28:16,121 INFO [ProcedureUtil] - (user1,67) END Search
bash fun.sh
休斯顿，我们有问题了。
控制台输出：
date: extra operand ‘+%H’
Try 'date --help' for more information.
date: extra operand ‘+%M’
Try 'date --help' for more information.
date: extra operand ‘+%S’
Try 'date --help' for more information.
date: extra operand ‘+%N’
Try 'date --help' for more information.
fun.sh: line 21: /1000000: syntax error: operand expected (error token is "/1000000")
应该是这样的
Begins: 2161626857
Ends: 2161626939
DIFFERENCE -----
The difference is 82 milliseconds
------
Begins: 216162343
Ends: 216162355
DIFFERENCE -----
The difference is 162 milliseconds
如果 s.log 是：
2019-02-22 06:27:06,857 INFO [ProcedureUtil] - (user1,14) START Search
2019-02-22 06:27:06,939 INFO [ProcedureUtil] - (user1,14) END Search
控制台输出：
Begins: 2161626857
Ends: 2161626939
DIFFERENCE -----
The difference is 82 milliseconds
-------
Begins: 2161626857
Ends: 2161626939
DIFFERENCE -----
The difference is 82 milliseconds
-------
两次一样，应该是一次。
求助我会很优雅
我正在考虑一个案例： 其中 s.log 是：
2019-02-22 06:27:06,857 INFO [ProcedureUtil] - (user1,14) START Search
2019-02-22 06:27:06,939 INFO [ProcedureUtil] - (user1,14) END Search
2019-02-22 07:28:16,088 INFO [ProcedureUtil] - (user1,67) START Split
2019-02-22 07:28:16,121 INFO [ProcedureUtil] - (user1,67) END Split
2019-02-22 07:28:16,088 INFO [ProcedureUtil] - (user1,67) START Search
2019-02-22 07:28:16,121 INFO [ProcedureUtil] - (user1,67) END Search
2019-02-25 20:59:59,999 INFO [ProcedureUtil] - (user1,17) START Search
2019-02-25 02:59:59,999 INFO [ProcedureUtil] - (user1,18) START Search
错误的解决方案：
COUNTER=0
while IFS= read -r line || [[ -n $line ]]; do
if [[ $line == *"START Search"* ]]; then
start=$(time2millis "$line")
echo "Begins: $start"
elif [[ $line == *"END Search"* ]]; then
end=$(time2millis "$line")
echo "Ends: $end"
# Assume every END has a preceding START
difference "$start" "$end"
else
COUNTER=$((COUNTER+1))
echo "$COUNTER"
fi
done < s.log
控制台输出：
Begins: 23226857
Ends: 23226939
DIFFERENCE -----
The difference is 82 milliseconds
------
Begins: 26896088
Ends: 26896121
DIFFERENCE -----
The difference is 33 milliseconds
------
Begins: 75599999
Begins: 10799999
顾虑：
- 你正在逐行阅读文件，但你没有在任何地方使用
$line：你正在 grep 整个文件 - 因为 grep 调用 returns 多次匹配，你的
$start变量包含换行符 因为你没有在
date调用中引用变量，所以看起来像这样date -d 06:27:06,857 07:28:16,088 '+%H'显然参数太多了 date
大多数错误都可以通过简化代码来消除。请注意，您的 start 和 end 函数在功能上是相同的。
difference() {
local diff=$(( - ))
printf 'DIFFERENCE -----\nThe difference is %d milliseconds\n------\n' "$diff"
}
time2millis() {
local time=$(echo "" | cut -d ' ' -f 2)
IFS=:, read -r hh mm ss nnn <<<"$time"
# be aware of invalid octal numbers 08 and 09:
# each component of the time must be handled as a decimal number
echo "$(( (((10#$hh) * 60 + 10#$mm) * 60 + 10#$ss) * 1000 + 10#$nnn ))"
}
while IFS= read -r line || [[ -n $line ]]; do
if [[ $line == *"START Search"* ]]; then
start=$(time2millis "$line")
echo "Begins: $start"
elif [[ $line == *"END Search"* ]]; then
end=$(time2millis "$line")
echo "Ends: $end"
# Assume every END has a preceding START
difference "$start" "$end"
fi
done < s.log
进一步发展：
这里不考虑END发生在第二天的情况
2019-02-28 23:59:59,999 blah blah START Search 2019-03-01 00:00:00,001 blah blah END Search您将看到 -86399998
的差异，而不是返回 2 毫秒的差异
- 您没有考虑夏令时（除非您的日志以 UTC 格式记录时间戳）。
我认为可以通过使用 date 解析日期和时间来解决这些问题：
time2millis() {
local epoch=$( date -d "$(echo "" | cut -d ' ' -f 1,2)" '+%s' )
local millis=0
[[ =~ ,([0-9]+) ]] && millis=${BASH_REMATCH[1]}
echo "$(( $epoch * 1000 + 10#$millis ))"
}
完整的解决方案：
difference() {
local diff=$(( - ))
while (( diff < 0 )); do (( diff += 86400000 )); done
printf "DIFFERENCE -----\nThe difference is %d milliseconds\n------\n" "$diff"
}
# returns milliseconds since 1970-01-01 00:00:00 UTC
time2millis() {
local epoch=$( date -d "$(echo "" | cut -d ' ' -f 1,2)" '+%s' )
local millis=0
[[ =~ ,([0-9]+) ]] && millis=${BASH_REMATCH[1]}
echo "$(( $epoch * 1000 + 10#$millis ))"
}
declare -A startTime startLine
while IFS= read -r line || [[ -n $line ]]; do
read -a words <<<"$line"
key=${words[5]}
if [[ "${words[6]} ${words[7]}" == "START Search" ]]; then
startLine[$key]=$line
startTime[$key]=$(time2millis "$line")
elif [[ "${words[6]} ${words[7]}" == "END Search" ]]; then
if [[ -z ${startTime[$key]} ]]; then
echo "END seen with no START: $line"
else
end=$(time2millis "$line")
difference "$key" "${startTime[$key]}" "$end"
unset startLine[$key] startTime[$key]
fi
fi
done < s.log
echo "Searches STARTed but not ENDed:"
printf "%s\n" "${startLine[@]}"
给定你的输入数据，输出
DIFFERENCE (user1,14) -----
The difference is 82 milliseconds
------
DIFFERENCE (user1,67) -----
The difference is 33 milliseconds
------
Searches STARTed but not ENDed:
2019-02-25 02:59:59,999 INFO [ProcedureUtil] - (user1,18) START Search
2019-02-25 20:59:59,999 INFO [ProcedureUtil] - (user1,17) START Search