可以执行
的gnuplot>使用1:2标题'数据A'绘制“fileA.dat”,\ “fileB.dat”使用1:3标题'数据B'
使用python gnuplot.py,以下函数可以正常工作:
def create_plot():
...
for target in TARGET_LIST:
for protocol in PROTOCOL_LIST:
input_file_name = "%s/%s.db" % (DATA_DIR, target)
shortname = input_file_name.split('/')[-1]
shortname = shortname.rstrip('.db')
input_list = read_lines(input_file_name)
write_plot_file_name = '%s/%s.%s.write.out' % (DATA_DIR, shortname, protocol)
write_plot_file = open(write_plot_file_name, 'w')
read_plot_file_name = '%s/%s.%s.read.out' % (DATA_DIR, shortname, protocol)
read_plot_file = open(read_plot_file_name, 'w')
ping_plot_file_name = '%s/%s.ping.out' % (DATA_DIR, shortname)
ping_plot_file = open(ping_plot_file_name, 'w')
for line in input_list[ limit: ]:
line = line.rstrip("\n")
line_protocol, line_verb, delta, timestamp = line.split('|')
if line_protocol == protocol:
if line_verb == 'write':
write_plot_file.write("%s,%s\n" % (timestamp, delta))
else:
read_plot_file.write("%s,%s\n" % (timestamp, delta))
elif line_protocol == 'ping':
ping_plot_file.write("%s,%s\n" % (timestamp, delta))
#gnuplot stuff
png_file = "%s/%s.%s.png" % (PLOT_DIR, shortname, protocol)
title = '%s history for %s ' % (protocol, shortname)
gnuplot = Gnuplot.Gnuplot()
gnuplot.title(title)
gnuplot('set style data linespoints')
gnuplot('set ylabel "seconds"')
gnuplot('set xlabel "month:day:hour:minute:seconds:millisecond"')
gnuplot('set xdata time')
gnuplot('set timefmt "%s"')
gnuplot('set format x "%m:%d:%H:%M:%S:%MS"')
gnuplot('set xtics nomirror rotate by -90')
gnuplot('set datafile separator ","')
gnuplot('set autoscale')
gnuplot('set grid xtics ytics')
gnuplot('set terminal png size 900,899')
gnuplot('set output "%s"' % (png_file))
cmd = ' "%s" using 1:2 axes x1y1 title "write" with lines, \
"%s" using 1:2 axes x1y1 title "read" with lines, \
"%s" using 1:2 axes x1y2 title "ping" with lines' % \
(write_plot_file_name, read_plot_file_name, ping_plot_file_name)
try:
gnuplot.plot(cmd)
except Error as why:
print "gnuplot choked: %s" % (why)
但是,当我做pythonic的事情并把它分成两个函数时:
def read_data():
#do the same file/data munging above
png_file = "%s/%s.%s.png" % (PLOT_DIR, shortname, protocol)
title = '%s history for %s ' % (protocol, shortname)
cmd = ' "%s" using 1:2 axes x1y1 title "write" with lines, \
"%s" using 1:2 axes x1y1 title "read" with lines, \
"%s" using 1:2 axes x1y2 title "ping" with lines' % \
(write_plot_file_name, read_plot_file_name, ping_plot_file_name)
def create_plot(png_file, title, cmd):
#call the gnuplot 'set' strings as above
gnuplot = Gnuplot.Gnuplot()
gnuplot.title(title)
gnuplot('set style data linespoints')
gnuplot('set ylabel "seconds"')
gnuplot('set xlabel "month:day:hour:minute:seconds:millisecond"')
gnuplot('set xdata time')
gnuplot('set timefmt "%s"')
gnuplot('set format x "%m:%d:%H:%M:%S:%MS"')
gnuplot('set xtics nomirror rotate by -90')
gnuplot('set datafile separator ","')
gnuplot('set autoscale')
gnuplot('set grid xtics ytics')
gnuplot('set terminal png size 900,899')
gnuplot('set output "%s"' % (png_file))`)
png_file = "%s/%s.%s.png" % (PLOT_DIR, shortname, protocol)
title = '%s history for %s ' % (protocol, shortname)
cmd = ' "%s" using 1:2 axes x1y1 title "write" with lines, \
"%s" using 1:2 axes x1y1 title "read" with lines, \
"%s" using 1:2 axes x1y2 title "ping" with lines' % \
(write_plot_file_name, read_plot_file_name, ping_plot_file_name)
def create_plot(png_file, title, cmd):
#call the gnuplot 'set' strings as above
gnuplot = Gnuplot.Gnuplot()
gnuplot.title(title)
gnuplot('set style data linespoints')
gnuplot('set ylabel "seconds"')
gnuplot('set xlabel "month:day:hour:minute:seconds:millisecond"')
gnuplot('set xdata time')
gnuplot('set timefmt "%s"')
gnuplot('set format x "%m:%d:%H:%M:%S:%MS"')
gnuplot('set xtics nomirror rotate by -90')
gnuplot('set datafile separator ","')
gnuplot('set autoscale')
gnuplot('set grid xtics ytics')
gnuplot('set terminal png size 900,899')
gnuplot('set output "%s"' % (png_file))`)
这会以任何数量有趣的方式失败,显然是因为gnuplot试图绘制字符串本身而不是像第一种情况那样解释它。
这里发生了什么?有没有办法解决这个问题?
答案 0 :(得分:0)
得到了它:
这两个函数都没有执行file.close(bad!bad!),这在单个函数的范围内是正常的 - 文件描述符仍然有效。
当
write_plot_file.close()
read_plot_file.close()
ping_plot_file.close()
包含在read_data()中一切都更好。