gnuplot 读取逗号分隔的数据文件

        有时，我们的数据文件中各个数据之间是用逗号作为分隔符的，比如标准的以“CSV”为后缀的那种数据文件。如果在逗号之后没有空格分隔，默认情况下gnuplot是无法直接读取的。

        这时可以有两种方案，第一种是提前处理一下数据文件，比如将逗号替换为空格，随便一个文本处理软件都能很轻松的做这种替换。但是有时我们有很多这样的数据文件，每个都这样处理一下也挺麻烦的。

        第二种方法就是在gnuplot中给出文件分隔符的信息，让gnuplot能够读懂我们的文件。下面将要说的就是这种方法。比如我们有如下的文件：

1. -3,0.1,0.0001234098  
2. -2.9,0.1062699256,0.0002226299  
3. -2.8,0.1131221719,0.000393669  
4. -2.7,0.1206272618,0.0006823281  
5. -2.6,0.1288659794,0.0011592292  
6. -2.5,0.1379310345,0.0019304541  
7. -2.4,0.1479289941,0.0031511116  
8. -2.3,0.1589825119,0.0050417603  
9. -2.2,0.1712328767,0.0079070541  
10. -2.1,0.1848428835,0.0121551783  
11. -2,0.2,0.0183156389  
12. -1.9,0.2169197397,0.0270518469  
13. -1.8,0.2358490566,0.0391638951  
14. -1.7,0.2570694087,0.0555762126  
15. -1.6,0.2808988764,0.0773047404  
16. -1.5,0.3076923077,0.1053992246  
17. -1.4,0.3378378378,0.1408584209  
18. -1.3,0.3717472119,0.184519524  
19. -1.2,0.4098360656,0.2369277587  
20. -1.1,0.4524886878,0.2981972794  
21. -1,0.5,0.3678794412  
22. -0.9,0.5524861878,0.4448580662  
23. -0.8,0.6097560976,0.527292424  
24. -0.7,0.6711409396,0.6126263942  
25. -0.6,0.7352941176,0.6976763261  
26. -0.5,0.8,0.7788007831  
27. -0.4,0.8620689655,0.852143789  
28. -0.3,0.9174311927,0.9139311853  
29. -0.2,0.9615384615,0.9607894392  
30. -0.1,0.9900990099,0.9900498337  
31. 0,1,1  
32. 0.1,0.9900990099,0.9900498337  
33. 0.2,0.9615384615,0.9607894392  
34. 0.3,0.9174311927,0.9139311853  
35. 0.4,0.8620689655,0.852143789  
36. 0.5,0.8,0.7788007831  
37. 0.6,0.7352941176,0.6976763261  
38. 0.7,0.6711409396,0.6126263942  
39. 0.8,0.6097560976,0.527292424  
40. 0.9,0.5524861878,0.4448580662  
41. 1,0.5,0.3678794412  
42. 1.1,0.4524886878,0.2981972794  
43. 1.2,0.4098360656,0.2369277587  
44. 1.3,0.3717472119,0.184519524  
45. 1.4,0.3378378378,0.1408584209  
46. 1.5,0.3076923077,0.1053992246  
47. 1.6,0.2808988764,0.0773047404  
48. 1.7,0.2570694087,0.0555762126  
49. 1.8,0.2358490566,0.0391638951  
50. 1.9,0.2169197397,0.0270518469  
51. 2,0.2,0.0183156389  
52. 2.1,0.1848428835,0.0121551783  
53. 2.2,0.1712328767,0.0079070541  
54. 2.3,0.1589825119,0.0050417603  
55. 2.4,0.1479289941,0.0031511116  
56. 2.5,0.1379310345,0.0019304541  
57. 2.6,0.1288659794,0.0011592292  
58. 2.7,0.1206272618,0.0006823281  
59. 2.8,0.1131221719,0.000393669  
60. 2.9,0.1062699256,0.0002226299  
61. 3,0.1,0.0001234098  

可以看到，数据有三列，用逗号来分隔，我们下面的例子中之用到前两列。如果直接用如下命令的话得到的不是我们希望的结果。

1. Plot 'sample.csv'  

gnuplot 只解析出了第一列的数据。如果我们告诉gnuplot我们的数据有两列会怎样呢？

1. Plot 'sample.csv' using 1:2  

gnuplot 会抱怨说：

1. gnuplot> plot 'sample.csv' using 1:2  
2.                                    ^  
3.          warning: Skipping data file with no valid points  
4.                                     ^  
5.          x range is invalid  

正确的方法是这样的：

plot 'sample.csv' using 1:2 "%lf,%lf"

        格式字符串的格式与C语言中scanf的格式字符串是类似的，实际上gnuplot最后就是用的scanf函数来读取数据。%lf表示按照double型浮点数类型来读取。需要注意的是gnuplot的格式化字符串不支持%f。