trouble transforming variables numeric

[F_user]

I worked on my Data in SPSS and saved it as csv. In SPSS all variables are type numeric. Now I am working on it in R and there are some variables that are somehow in the wrong format.

mean(PISA$MATH, na.rm = TRUE)
[1] NA
Warning message:
In mean.default(PISA$MATH, na.rm = TRUE) :
Argument ist weder numerisch noch boolesch: gebe NA zurück
It says it`s not numeric.

However if I use the as.numeric function I loose a huge amount of data:

Warning message:
NAs introduced by coercion

Is there an option to transform the data to numeric without introducing NAs? Or does anyone know if I made a mistake in the process from SPSS to R?

[yiy]

@F_user
Hi there,
Did you load your original data into R from a CSV file OR a PASS file?
Would you like to provide a snippet of your code here to show us how you loaded your source data into R?

[F_user]

I loaded a CSV file into r:

PISA <- read.table("PISA.csv", header=TRUE, sep=";", dec=".")

[M_AcostaCH]

Hi @F_user, for put all columns as numeric you could try in this form:

data <- read.csv("yourfile.csv", stringsAsFactors = FALSE)

data <- sapply(data, as.numeric)

But as.numeric() return NA for any values that cannot be converted to numeric, such as character strings.
To replace NA values with zero, you can use the is.na().

data[is.na(data)] <- 0

[F_user]

Thank you.

I tried it and I get the error " more column than colums names"

[M_AcostaCH]

Well, but for get a better help try to put reproducible example for better understand the situation

Reprex

Othe way is paste the result of dput():

data <- read.csv("yourfile.csv", stringsAsFactors = FALSE) #load data
dput(head(data,20)) # For get all columns and the firs 20 rows of the data

[nirgrahamuk]

@F_user

I worked on my Data in SPSS and saved it as csv.

if you were to use the haven package , you could presumably directly read your SPSS file, without bothering with all the issues csv's can introduce.
Import and Export SPSS, Stata and SAS Files • haven (tidyverse.org)