我需要评估矩阵中每个元素的条件，其中测试涉及dimnames每个索引的 。具体来说，如果元素的rowname和colname共享特定功能，我想将该元素设置为零。

m <- matrix(rnorm(81),nrow=9)
colnames(m) <- paste(c(rep("A",3),rep("B",3),rep("C",3)),1:ncol(m),sep = "_")
rownames(m) <- paste(c(rep("A",3),rep("B",3),rep("C",3)),1:nrow(m),sep = "_")
m

            A_1        A_2         A_3          B_4        B_5         B_6        C_7        C_8         C_9
A_1 -0.03201198 -2.2241923 -0.65584334 -0.346745371 -0.9263060 -1.99181830  0.9138187 -2.4959751 -0.96723090
A_2 -1.44319826 -0.2225057 -1.35327091 -0.009194619  0.5798469  2.42753826 -1.4574564 -0.8858597 -1.41595891
A_3 -0.05863965 -0.2177708 -0.39131739  0.729532751  1.4106448  0.15899085 -1.7521345  0.5398222 -0.05073061
B_4  1.11006840  1.0315201  0.10434758 -0.508430234 -1.7095192  0.90913528  1.7367210 -0.9006098 -1.41698688
B_5  0.21405173  0.4735690  0.42655214 -0.367748304  0.9820261 -0.77933908  1.1326391  0.5316226  2.24951820
B_6  0.27153476 -0.3506076  0.16943749 -0.666135969  0.2962018  1.12236640 -1.3103133 -1.9494454 -0.57526358
C_7  1.69732641 -1.1439368 -0.02734925 -0.814635435  0.6658583  0.68069434  0.3330596 -1.2564933  0.15807742
C_8  0.35194835 -0.7075880 -0.45814046  0.773997223 -0.6530986  0.01295098  0.2557955  1.4658751 -3.33651509
C_9  0.58610083  0.7908394 -1.38909037 -0.742739398 -0.3745243  2.80990368  0.2172529 -0.3672324  0.56309688

我想要的结果可以通过 for 循环轻松实现，但是在处理大型矩阵时，这当然非常慢。

for (i in 1:nrow(m)) {
  for (j in 1:ncol(m)) {
    m[i,j] <- ifelse(sub("_.*", "1", rownames(m)[i])==sub("_.*", "1", colnames(m)[j]),0,m[i,j])
  }
}
m

          A_1        A_2         A_3          B_4        B_5         B_6        C_7        C_8         C_9
A_1 0.0000000  0.0000000  0.00000000 -0.346745371 -0.9263060 -1.99181830  0.9138187 -2.4959751 -0.96723090
A_2 0.0000000  0.0000000  0.00000000 -0.009194619  0.5798469  2.42753826 -1.4574564 -0.8858597 -1.41595891
A_3 0.0000000  0.0000000  0.00000000  0.729532751  1.4106448  0.15899085 -1.7521345  0.5398222 -0.05073061
B_4 1.1100684  1.0315201  0.10434758  0.000000000  0.0000000  0.00000000  1.7367210 -0.9006098 -1.41698688
B_5 0.2140517  0.4735690  0.42655214  0.000000000  0.0000000  0.00000000  1.1326391  0.5316226  2.24951820
B_6 0.2715348 -0.3506076  0.16943749  0.000000000  0.0000000  0.00000000 -1.3103133 -1.9494454 -0.57526358
C_7 1.6973264 -1.1439368 -0.02734925 -0.814635435  0.6658583  0.68069434  0.0000000  0.0000000  0.00000000
C_8 0.3519484 -0.7075880 -0.45814046  0.773997223 -0.6530986  0.01295098  0.0000000  0.0000000  0.00000000
C_9 0.5861008  0.7908394 -1.38909037 -0.742739398 -0.3745243  2.80990368  0.0000000  0.0000000  0.00000000

概念上类似于这个问题，但我不知道如何评估家庭dimnames内的条件apply。有什么建议？谢谢！

回答

rnm <- sub("_.*", "1", rownames(m))
cnm <- sub("_.*", "1", colnames(m))
m[rnm[row(m)] == cnm[col(m)]] <- 0

 m
           A_1         A_2         A_3         B_4        B_5         B_6        C_7        C_8         C_9
A_1  0.0000000  0.00000000  0.00000000 -2.84612856  0.1611173  0.29004403 -0.8844186  0.8363131 -0.57395543
A_2  0.0000000  0.00000000  0.00000000  1.92166045  0.2671856 -0.50424582  0.8672366 -2.2496354  0.04046654
A_3  0.0000000  0.00000000  0.00000000 -0.01515657  0.6775447 -0.03862803  1.7764642  0.7146040  0.33652933
B_4  1.0842025  0.32981334 -0.61961179  0.00000000  0.0000000  0.00000000  0.3366096  0.5196087  1.67367867
B_5 -1.2897726  0.04082185  1.62661008  0.00000000  0.0000000  0.00000000 -1.6750488  0.5464289  0.98246881
B_6 -0.4213025 -0.70439232  0.09091241  0.00000000  0.0000000  0.00000000  0.7050487 -0.4151445 -0.04604658
C_7  1.0594241 -1.43071282 -0.75394573  1.08360040  1.3646551 -0.88687658  0.0000000  0.0000000  0.00000000
C_8  1.3908082  1.12093371  1.73690687  1.05202987  0.6152715  0.45601621  0.0000000  0.0000000  0.00000000
C_9 -0.3026523  1.14507291 -1.04714611 -2.38087279 -0.6976168 -0.96394113  0.0000000  0.0000000  0.00000000

xtabs(Freq ~ Var1 + Var2, transform(as.data.frame.table(m), 
     Freq = Freq * (substr(Var1, 1, 1) != substr(Var2, 1, 1))))