Gnuplot：每分鐘頻率

我有一個包含 1000 行的示例日志檔案，看起來像這樣，

TIME,STATUS
09:00,OK
09:00,TEMP
09:00,TEMP
09:00,TEMP
09:00,TEMP
09:00,TEMP
09:01,OK
09:01,OK
09:01,OK
09:01,PERM
09:01,TEMP
09:01,TEMP
09:02,OK
09:02,TEMP
09:02,TEMP
09:03,OK
09:03,PERM
09:03,PERM
09:03,TEMP
09:03,TEMP
09:04,OK
09:04,PERM
09:04,PERM
09:04,TEMP
09:04,TEMP
09:04,TEMP
09:05,OK
09:05,OK
09:05,OK
09:05,PERM
09:05,TEMP
09:05,TEMP
09:05,TEMP
09:05,TEMP
09:06,OK
09:06,OK
09:06,PERM
09:06,PERM
09:06,PERM
09:06,PERM
09:06,TEMP
09:06,TEMP
09:06,TEMP
09:06,TEMP
09:06,TEMP
09:07,OK
09:07,OK
09:07,TEMP
09:07,TEMP
09:07,TEMP
09:08,OK
09:08,OK
09:08,OK
09:08,OK
09:08,OK
09:08,OK
09:08,OK
09:08,TEMP
09:08,TEMP
09:08,TEMP
09:08,TEMP
09:09,OK
09:09,OK
09:09,OK
09:09,PERM
09:10,OK
09:10,PERM
09:10,PERM
09:10,TEMP
09:11,OK
09:11,OK
09:11,OK
09:11,OK
09:11,PERM
09:11,PERM
09:11,PERM
09:11,PERM
09:11,TEMP
09:11,TEMP
09:11,TEMP
09:12,PERM
09:12,TEMP
09:12,TEMP
09:13,OK
09:13,OK
09:13,OK
09:13,OK
09:13,OK
09:13,PERM
09:13,PERM
09:13,PERM
09:13,TEMP
09:13,TEMP
09:14,OK
09:14,OK
09:14,OK
09:14,PERM
09:14,PERM
09:14,PERM
09:14,PERM
09:14,TEMP
09:16,OK
09:16,OK
09:16,OK
09:16,PERM
09:16,PERM
09:16,TEMP
09:16,TEMP
09:17,OK
09:17,OK
09:17,PERM
09:17,PERM
09:18,OK
09:18,OK
09:18,OK
09:18,OK
09:18,OK
09:18,PERM
09:18,PERM
09:18,TEMP
09:18,TEMP
09:18,TEMP
09:19,OK
09:19,OK
09:19,OK
09:19,OK
09:19,OK
09:19,PERM
09:20,OK
09:20,OK
09:20,PERM
09:20,PERM
09:20,TEMP
09:20,TEMP
09:21,OK
09:21,OK
09:21,OK
09:21,PERM
09:21,TEMP
09:22,OK
09:22,OK
09:22,PERM
09:22,PERM
09:22,TEMP
09:22,TEMP
09:23,OK
09:23,PERM
09:23,PERM
09:23,PERM
09:23,TEMP
09:23,TEMP
09:23,TEMP
09:24,PERM
09:24,PERM
09:24,PERM
09:25,OK
09:25,OK
09:25,PERM
09:25,TEMP
09:26,OK
09:26,OK
09:26,OK
09:26,OK
09:26,OK
09:26,PERM
09:26,TEMP
09:27,OK
09:27,OK
09:27,OK
09:27,PERM
09:27,PERM
09:27,TEMP
09:27,TEMP
09:27,TEMP
09:28,PERM
09:28,PERM
09:28,PERM
09:28,PERM
09:29,OK
...

而最終檔案在同一時間范圍內將有 10K 行。我需要創建一個圖表來顯示每分鐘 TEMP、PERM 和 OK 的狀態數。所以我想用一條線表示狀態（TEMP、PERM 和 OK），在 X 軸上繪制時間，在 Y 軸上繪制出現頻率。

我僅在 2 天前從標準 repo 在我的 Ubuntu 20.04.4 LTS 上安裝了 Gnuplot：

bi@green:bin$ apt list gnuplot* 2>/dev/null | grep installed
gnuplot-data/focal,focal,now 5.2.8 dfsg1-2 all [installed,automatic]
gnuplot-qt/focal,now 5.2.8 dfsg1-2 amd64 [installed,automatic]
gnuplot/focal,focal,now 5.2.8 dfsg1-2 all [installed]

到目前為止，我還沒有做到更多，

#!/bin/bash
x=logoutcol
cat $x

gnuplot -p <<-EOF
#set ytics scale 0
#set yzeroaxis
reset
set format x "%H:%M" time
set xdata time

set yrange [0:*]
set ylabel "Occurences"
set ytics 2

#set margin at screen 0.95

binwidth=60
bin(val) = binwidth * floor(val/binwidth)

set boxwidth binwidth

set datafile separator ","
set term png
set output "$x.png"
plot "$x" using (bin(timecolumn(1,"%H%M"))):(2) smooth freq with boxes
EOF

shotwell $x.png
rm $x.png

產生這個：

任何幫助都感激不盡。

uj5u.com熱心網友回復：

我很確定在 SO 上有一個幾乎相同的問題，但是，我似乎找不到它，可能是因為我無法為 SO 的搜索功能找到正確的關鍵字。

關鍵是布爾運算式(strcol(2) eq word(myKeys,i))和smooth frequency. 如果第二列的值與您的關鍵字相同，則運算式結果為1，0否則。您不需要像創建其他直方圖那樣的 bin，因為您需要 1 分鐘的 bin（并且您的時間解析度已經是 1 分鐘）。

檢查以下示例作為進一步優化的起點。

腳本：

### count occurrences of keywords
reset session

$Data <<EOD
# TIME,STATUS
09:00,OK
09:00,TEMP
09:00,TEMP
09:00,TEMP
09:00,TEMP
09:00,TEMP
09:01,OK
09:01,OK
09:01,OK
09:01,PERM
09:01,TEMP
09:01,TEMP
09:02,OK
09:02,TEMP
09:02,TEMP
09:03,OK
09:03,PERM
09:03,PERM
09:03,TEMP
09:03,TEMP
09:04,OK
09:04,PERM
09:04,PERM
09:04,TEMP
09:04,TEMP
09:04,TEMP
09:05,OK
09:05,OK
09:05,OK
09:05,PERM
09:05,TEMP
09:05,TEMP
09:05,TEMP
09:05,TEMP
09:06,OK
09:06,OK
09:06,PERM
09:06,PERM
09:06,PERM
09:06,PERM
09:06,TEMP
09:06,TEMP
09:06,TEMP
09:06,TEMP
09:06,TEMP
09:07,OK
09:07,OK
09:07,TEMP
09:07,TEMP
09:07,TEMP
09:08,OK
09:08,OK
09:08,OK
09:08,OK
09:08,OK
09:08,OK
09:08,OK
09:08,TEMP
09:08,TEMP
09:08,TEMP
09:08,TEMP
09:09,OK
09:09,OK
09:09,OK
09:09,PERM
09:10,OK
09:10,PERM
09:10,PERM
09:10,TEMP
09:11,OK
09:11,OK
09:11,OK
09:11,OK
09:11,PERM
09:11,PERM
09:11,PERM
09:11,PERM
09:11,TEMP
09:11,TEMP
09:11,TEMP
09:12,PERM
09:12,TEMP
09:12,TEMP
09:13,OK
09:13,OK
09:13,OK
09:13,OK
09:13,OK
09:13,PERM
09:13,PERM
09:13,PERM
09:13,TEMP
09:13,TEMP
09:14,OK
09:14,OK
09:14,OK
09:14,PERM
09:14,PERM
09:14,PERM
09:14,PERM
09:14,TEMP
09:16,OK
09:16,OK
09:16,OK
09:16,PERM
09:16,PERM
09:16,TEMP
09:16,TEMP
09:17,OK
09:17,OK
09:17,PERM
09:17,PERM
09:18,OK
09:18,OK
09:18,OK
09:18,OK
09:18,OK
09:18,PERM
09:18,PERM
09:18,TEMP
09:18,TEMP
09:18,TEMP
09:19,OK
09:19,OK
09:19,OK
09:19,OK
09:19,OK
09:19,PERM
09:20,OK
09:20,OK
09:20,PERM
09:20,PERM
09:20,TEMP
09:20,TEMP
09:21,OK
09:21,OK
09:21,OK
09:21,PERM
09:21,TEMP
09:22,OK
09:22,OK
09:22,PERM
09:22,PERM
09:22,TEMP
09:22,TEMP
09:23,OK
09:23,PERM
09:23,PERM
09:23,PERM
09:23,TEMP
09:23,TEMP
09:23,TEMP
09:24,PERM
09:24,PERM
09:24,PERM
09:25,OK
09:25,OK
09:25,PERM
09:25,TEMP
09:26,OK
09:26,OK
09:26,OK
09:26,OK
09:26,OK
09:26,PERM
09:26,TEMP
09:27,OK
09:27,OK
09:27,OK
09:27,PERM
09:27,PERM
09:27,TEMP
09:27,TEMP
09:27,TEMP
09:28,PERM
09:28,PERM
09:28,PERM
09:28,PERM
09:29,OK
EOD

set datafile separator comma
myKeys = "OK TEMP PERM"
myKey(i) = word(myKeys,i)

myTimeFmt = "%H:%M"
set format x myTimeFmt timedate

plot for [i=1:words(myKeys)] $Data u (timecolumn(1,myTimeFmt)):(strcol(2) eq word(myKeys,i)) smooth freq w lp pt 7 ti word(myKeys,i)
### end of script

結果：

標籤：重击 gnuplot

上一篇：從bash選單執行一個選項

下一篇：使用帶有特殊字符的sed更改一行文本