你的group是字符型，你要把后面八个row的.换成空格就行了。.是代表数值型空缺的

但换成空格就出现这些错误了啊
WARNING: CLASS=12.40 has only 1 value in DATA= data set or BY group.
WARNING: Within-class covariance matrix cannot be evaluated for CLASS=12.40 in DATA= data set or BY group; the pooled covariance
         matrix will be used.
WARNING: CLASS=15.00 has only 1 value in DATA= data set or BY group.
WARNING: Within-class covariance matrix cannot be evaluated for CLASS=15.00 in DATA= data set or BY group; the pooled covariance
         matrix will be used.
WARNING: CLASS=28.60 has only 1 value in DATA= data set or BY group.
WARNING: Within-class covariance matrix cannot be evaluated for CLASS=28.60 in DATA= data set or BY group; the pooled covariance
         matrix will be used.
WARNING: CLASS=7.90 has only 1 value in DATA= data set or BY group.
WARNING: Within-class covariance matrix cannot be evaluated for CLASS=7.90 in DATA= data set or BY group; the pooled covariance

读取的观测数 18
使用的观测数 10
你上边的程序程序没问题，SAS警告也没问题，一切正常，我运行了
data d511;
    input x1-x4 group ;
    cards;
13.85 2.79  7.80 49.60 1
22.31 4.67 12.31 47.80 1
28.82 4.63 16.18 62.15 1
15.29 3.54  7.50 43.20 1
28.79 4.90 16.12 58.10 1
2.18 1.06  1.22 20.60 2
3.85 0.80  4.06 47.10 2
11.40 0.00  3.50  0.00 2
3.66 2.42  2.14 15.10 2
12.10 0.00  5.68  0.00 2
8.85 3.38  5.17 26.10 .
28.60 2.40  1.20 127.0 .
20.70 6.70  7.60 30.20 .
7.90 2.40  4.30 33.20 .
3.19 3.20  1.43  9.90 .
12.40 5.10  4.43 24.60 .
16.80 3.40  2.31 31.30 .
15.00 2.70  5.02 64.00 .
;
* proc print data=d511;
run;
proc discrim data=d511 outstat=out simple wcov pcov
              wsscp psscp distance  anova manova
              pool=test list listerr posterr;
    class group;
    var x1-x4;
    priors prop;
run;

日志

NOTE: Updated analytical products:

      SAS/STAT 13.2
      SAS/ETS 13.2

NOTE: Additional host information:

X64_8PRO WIN 6.2.9200  Workstation

NOTE: “SAS 初始化”所用时间:
      实际时间          4.74 秒
      CPU 时间          1.40 秒

1    data d511;
2        input x1-x4 group ;
3        cards;

NOTE: 数据集 WORK.D511 有 18 个观测和 5 个变量。
NOTE: “DATA 语句”所用时间（总处理时间）:
      实际时间          0.16 秒
      CPU 时间          0.04 秒


22    ;
23   * proc print data=d511;
24    run;
25    proc discrim data=d511 outstat=out simple wcov pcov
26                 wsscp psscp distance  anova manova
27                 pool=test list listerr posterr;
28       class group;
29       var x1-x4;
30       priors prop;
31    run;

NOTE: 正在写入 HTML Body（主体）文件: sashtml.htm
WARNING: 8 observations in DATA= data set or BY group are excluded from the development of the classification criterion
         due to missing values on CLASS variable.
WARNING: A total of 8 of the 18 observations in the DATA= data set were excluded (frequency sum = 10).
NOTE: The posterior probability error rate estimate is less than zero for DATA= data set.  This is usually due to large
      discrepancies between prior probabilities and class sizes.
NOTE: 从数据集 WORK.D511. 读取了 18 个观测
NOTE: 数据集 WORK.OUT 有 93 个观测和 7 个变量。
NOTE: “PROCEDURE DISCRIM”所用时间（总处理时间）:
      实际时间          1.14 秒
      CPU 时间          0.54 秒

警告是正常的，不是错误。估计是把待判数据和分类数据写在一起了，这样方便，不过有些警告。实在在意的话，可以把后面几个待判观测数据写到Testdata中，原来的程序更省事。

哦哦，谢谢