从is.nan返回错误的逻辑(Returning incorrect logical from is.nan)

 我正在分析来自公开数据库的微阵列数据，我的输出是一个对数比矩阵，将实验数据与对照进行比较。 我发现我的一个列填充了NaN值，因为这种处理没有相应的控制。 我的数据框（ fc_initial ）尺寸为9049 x 647。 
 
> dim(fc_initial)
[1] 9049  647
 
 由NaN填充的列是第580列 
 
head(fc_initial[,580])
# [1] NaN NaN NaN NaN NaN NaN

summary(is.nan(fc_initial[,580]))
#    Mode    TRUE    NA's 
# logical    9049       0 
 
 当我尝试使用apply检查所有列时，R错误地返回此列的所有FALSE 。 
 
fc_num <- apply(fc_initial, 2, is.nan)
summary(fc_num[ ,580])
#    Mode   FALSE    NA's 
# logical    9049       0 
 
 我已经使用fc_initial和fc_num视图在视觉上确认了这fc_num ，我已经仔细检查过我正在寻找使用which()的正确列。 
 
which(fc_num, arr.ind = TRUE)
#  row col
 
 which返回显示为TRUE单元格。 
 
 我错过了什么？ 为什么R为一列我在视觉上确认为TRUE的值返回一列FALSE ？ 

I am analyzing microarray data from a publicly available database, and my output is a log-ratio matrix comparing experimental data to controls. I have discovered that one of my columns is populated with NaN values due to the fact that this treatment does not have a corresponding control. My data frame's (fc_initial) dimensions are 9049 x 647.
 
> dim(fc_initial)
[1] 9049  647
 
the column that is populated by NaN is the 580th column
 
head(fc_initial[,580])
# [1] NaN NaN NaN NaN NaN NaN

summary(is.nan(fc_initial[,580]))
#    Mode    TRUE    NA's 
# logical    9049       0 
 
When I try to examine this for all columns using apply, R incorrectly returns all FALSE for this column.
 
fc_num <- apply(fc_initial, 2, is.nan)
summary(fc_num[ ,580])
#    Mode   FALSE    NA's 
# logical    9049       0 
 
I have visually confirmed this using view for both fc_initial and fc_num, and I have double checked that I am looking for the right column using which().
 
which(fc_num, arr.ind = TRUE)
#  row col
 
which returns no cells showing as TRUE.
 
What am I missing? Why is R returning a column of FALSE for a column of values I have visually confirmed to be TRUE?

原文：https://stackoverflow.com/questions/36895663