查找矩阵中重复次数最多的行

本文介绍了查找矩阵中重复次数最多的行的处理方法，对大家解决问题具有一定的参考价值，需要的朋友们下面随着小编来一起学习吧！ 问题描述

我在一个矩阵中有大约 10000 个样本的重复.我的矩阵有 1000 行和 6 列.列中的数字范围为 0:58，具体取决于样本.

I have about 10000 replicates of a sample in a matrix. My matrix has 1000 rows and 6 columns. Numbers in the columns range from 0:58 depending on the sample.

actual.prob <- c(.14, .14, .16, .13, .19, .24) million.rep <- replicate(10000, sample(1:6, 58, replace= T, actual.prob)) new.matrix <- matrix(nrow= 10000, ncol=6) for(i in 1:10000){ new.matrix[i,] <- as.vector(table(factor(million.rep[,i], levels=1:6))) } new.matrix[1:10,] [,1] [,2] [,3] [,4] [,5] [,6] [1,] 3 7 11 11 11 15 [2,] 7 6 12 5 19 9 [3,] 12 7 6 8 11 14 [4,] 6 7 16 6 11 12 [5,] 5 9 12 5 14 13 [6,] 9 4 14 7 10 14 [7,] 6 9 9 6 15 13 [8,] 9 4 8 8 11 18 [9,] 6 11 7 5 12 17 [10,] 7 6 9 9 15 12

我想查找是否有重复的样本.我尝试了 replicated() ，它告诉我哪些行是复制的，但我想查看该行而不必手动返回.有什么建议吗?

I want to find if any samples are repeated. I tried replicated() which tells me what rows are replicates, but I want to view the row without having to go back manually. Any suggestions?

推荐答案

这是一个data.table的实现:

library(data.table) dt <- data.table(new.matrix) head(dt[, list(repeats=.N, id=.I[[1]]), by=names(dt)][order(repeats, decreasing=T)], 20) # V1 V2 V3 V4 V5 V6 repeats id # 1: 5 7 11 8 13 14 4 543 # 2: 5 11 13 5 10 14 4 579 # 3: 6 8 6 10 12 16 4 1433 # 4: 6 9 9 9 9 16 4 1688 # 5: 8 8 9 7 10 16 4 2382 # 6: 6 10 8 7 11 16 4 2965 # 7: 7 9 11 8 11 12 4 3114 # 8: 8 8 10 7 10 15 4 3546 # 9: 7 8 12 9 9 13 4 5759 # 10: 7 7 13 9 10 12 4 9021 # 11: 8 10 8 8 12 12 3 81 # 12: 9 10 7 7 11 14 3 110 # 13: 7 11 8 6 12 14 3 130 # 14: 11 9 7 7 9 15 3 143 # 15: 8 10 10 7 11 12 3 330 # 16: 8 9 10 8 13 10 3 480 # 17: 9 10 7 10 11 11 3 542 # 18: 8 6 11 9 11 13 3 555 # 19: 7 10 7 6 10 18 3 577 # 20: 7 8 10 5 12 16 3 601

其中 repeats 是一行出现的次数，id 是矩阵中与该模式匹配的第一行.

where repeats is how many times a row shows up, and id the first row in the matrix that matches that pattern.