[R-meta] Dataframe is character

[Ken Beath]

More correctly you believe that your Excel file is all numeric and it obviously isn’t. Reading Excell files can be a bit nasty because it is easy in the process of editing to have cells that are formatted as text. This can even be rows after teh data.

You have 2 choices. Either fix the Excel file by deleting any extra rows and making sure everything is formatted as numeric  or save it as a csv and then use read.csv. I’ve found the second option much more reliable.

Ken

> On 17 Nov 2017, at 9:14 pm, P. Roberto Bakker <robertobakker at gmail.com> wrote:
> 
> I resend my mail as I was not member of this list - now I am.
> Roberto
> 
> ---------- Forwarded message ----------
> From: P. Roberto Bakker <robertobakker at gmail.com>
> Date: 2017-11-17 10:41 GMT+01:00
> Subject: Dataframe is character
> To: r-help at r-project.org
> 
> 
> 
> Hi everybody,
> 
> Question: why are my dataframe and numeric variables a character?
> 
> I read an excel file via readxl but my dataframe is a character, and
> numeric variables, eg "yi", are also a character.
> My excelfile is in English numeric
> Sometimes the dataframe was indeed a dataframe, but I do not know why it
> did sometimes.
> Thank you in advance, Roberto
> PS I used "guess". The problem is not solve by using "text", "numeric" etc
> 
> My syntax (I think I cannot send the excel file as binary?)
> 
>> library(readxl)
>> library(readxl)
>> library(metafor)
>> setwd("C:/docs/Work2/Statistic_Analyses/MetaQTcAD")
>> getwd()
> [1] "C:/docs/Work2/Statistic_Analyses/MetaQTcAD"
>> 
> 
>> dat <- read_excel("Hedges-g_QTc MA_R05.xlsx", sheet = 2, col_names=TRUE,
> col_types = c("guess"))
>> class("dat")
> [1] "character"
>> class("yi")
> [1] "character"
>> 
> 
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> R-sig-meta-analysis mailing list
> R-sig-meta-analysis at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-meta-analysis
>