Error in if (length(df) == 7 & unique(sapply(df, class)[5:7]) == "integer") { : the condition has length > 1

[HuangRuocheng]

Hi,

I have the error shown in the title. It was thown by tabix2dt in chrY_JH584300_random,

in which I used applyTbxByChr to read methDiffDB, because in WGBS it met core dump occasionally.

I checked chrY_JH584300_random methUnite data:

it has only 4 CpG sites, all of which Beta value is converted to integers;

then in some where of tabix2dt generated the following data.frame:

which will then be checked whether or not the 5-7 columns have the same type of class integer,

and then throw an error because 5th and 7th column have different type, leading to unique( FALSE, TRUE) and the condition has length > 1

My question is:

What dose 5th to 7th columns mean? Could I skip this check?

Thank you very much!

[alexg9010]

Hi @HuangRuocheng,

thanks for reporting this issue. Most likely, this error originates from the .setMethylDBNames function, which we use to determine the type of object based on the data. These checks seem to cause the problem:

methylKit/R/methylDBClasses.R

Lines 18 to 27 in ecf8584

	if( length(df) == 7 & unique(sapply(df,class)[5:7])=="integer"){
	setnames(x = df,old = names(df),
	new = c("chr","start","end","strand",
	"coverage","numCs","numTs"))
	} else if( length(df) == 7 & unique(sapply(df,class)[5:7])=="numeric"){
	setnames(x = df,old = names(df),
	new = c("chr","start","end","strand",
	"pvalue","qvalue","meth.diff"))

A simple fix for your problem would be to set the methylDBclass directly when calling .setMethylDBNames, then this check will be skipped.

Best,
Alex

[HuangRuocheng]

Thank you very much !!

As you said, I added " methylKit:::.setMethylDBNames(obj, methylDBclass = "methylDiffDB") " to the function which applyTbxByChr called, and it run smoothly!

Thanks again!