Question

如果我试图从Kingk中找到平均c(NA、NA、3、4、5、6、7、8、9、10),则我获得的数值是5.2,而不是预期的6.5。

# prepare data and write to file
write.table(data.frame(col1 = c(NA, NA, 3:10)),
        "my.na.txt", row.names = FALSE)

mean(c(NA, NA, 3:10), na.rm = TRUE) # 6.5

my.na <- read.csv.sql("my.na.txt", sep = " ",
        sql = "SELECT AVG(col1) FROM file") # 5.2

# this is identical to
sum(3:10)/10

unlink("my.na.txt") # remove file

这使我相信,ql(df)将国民价值视为零。能否在Kall电话中忽略(不包括)NA值,因为可以使用na.rm的论点(在R)进行?

Answer 1

修改你关于忽视<条码>NA数值的询问:

SELECT AVG(col1)
FROM file
WHERE col1 IS NOT "NA"

Answer 2

The problem is that the read.csv.sql function does not recognize the missing values, and converts them to zero, instead of NULL. This does not happen if you first load the data into a data.frame, and only then call sqldf.

d <- read.csv("my.na.txt")
sqldf("SELECT AVG(col1) FROM d") # 6.5

Answer 3

供今后参考。我发现,在研究这一问题时,可以找到解决办法。当被转往别处时,有色人种被转换为NUL,然后,你可以利用ISNUL或IS NOT NUL将他们排除在外。

友情链接