删除任何/所有 NaN 值的行/列

Question

在带有 pandas 的 python 中，我可以执行以下操作：

# Drop columns with ANY missing values
df2 = df.dropna(axis=1, how="any")

# Drop columns with ALL missing values
df2 = df.dropna(axis=1, how="all")

# Drop rows with ANY missing values
df2 = df.dropna(axis=0, how="any")

# Drop rows with ALL missing values
df2 = df.dropna(axis=0, how="all")

我如何类似地过滤 R data.table 中的行/列？

Answer 1

我们可以将Reduce 与| 或& 一起使用

library(data.table)
#Drop rows with any missing values
setDT(df1)[df1[, !Reduce(`|`, lapply(.SD, is.na))]]
#Drop rows with all missing values 
setDT(df1)[df1[, !Reduce(`&`, lapply(.SD, is.na))]]

#Drop columns with any and all missing values
Filter(function(x) !any(is.na(x)), df1)
Filter(function(x) !all(is.na(x)), df1)
#or use
setDT(df1)[, unlist(df1[, lapply(.SD, function(x) any(!is.na(x)))]), with = FALSE]
setDT(df1)[, unlist(df1[, lapply(.SD, function(x) all(!is.na(x)))]), with = FALSE]      

数据

set.seed(24)
df1 <- as.data.table(matrix(sample(c(NA, 0:5), 4*5, replace=TRUE), ncol=4))
df1[3] <- NA