R/extractcodesfunctions.R
In niekverw/ukpheno: CreateUKBiobankPhentoypes
Defines functions
Query_TSAmin_masterset
CheckIfColumnsAreAvailable
Query_Masterset2
Death_Masterset
Outcome_HES
#' @export
Outcome_HES<-function(dfmaster_SQL_merge,StrTrait,StrDescription,VctCodes,epidurfilter=0,StrColumnForHescodes,StrColumnnameForEventDate,StrColumnnameForVisitDate,StataOutputFile,
grep_with_wildcard=TRUE #recurrence_is_possible=TRUE # not for self reported data..
){
# epidurfilter=0
# dfmaster_SQL_merge=dfgpclinical
# dfmaster_SQL_merge=dfhesintables
### CHECK IF FILE EXISTS:
if( file.exists(StataOutputFile )) {
print(paste(StataOutputFile," ... already exists!, skipping"))
return(0)
}
print(paste("extracting outcomes from episode data, checking codes:",StrTrait,":",StrDescription,"-",StrColumnnameForVisitDate, "--",paste(StrColumnForHescodes,collapse=",") ) )
DiagnosisVars=c("n_eid","HXn","HXd","HXt","HXto","FUn","FUd","FUt","FUto") ### variables you wil create and you want to keep later...
### FUt/ FUto = NOT FOR GP DATA, only HESIN
#### RENAME COLUMNS
names(dfmaster_SQL_merge)[names(dfmaster_SQL_merge)==StrColumnnameForEventDate]<-"event_date"
names(dfmaster_SQL_merge)[names(dfmaster_SQL_merge)==StrColumnnameForVisitDate]<-"visit_date"
## remove records that have no visit date (relevant for ts_53_1_0 , ts_53_2_0).
dfmaster_SQL_merge_tmp<-dfmaster_SQL_merge[!is.na(dfmaster_SQL_merge$visit_date),]
########### GREP CODE AT THE COLUMNS 'StrColumnForHescodes' ;; TODO grep should be ^: paste(sep="","^",VctCodes, collapse='|')
#grepoper=c()
#for (i in StrColumnForHescodes ){
# grepoper<-c(grepoper,grep(paste(VctCodes, collapse='|'), dfmaster_SQL_merge_tmp[,i], ignore.case=FALSE))
#}
if (length(StrColumnForHescodes)==1){StrColumnForHescodes = c(StrColumnForHescodes,StrColumnForHescodes)}
if(grep_with_wildcard==TRUE){
grepoper <-unlist( mclapply( dfmaster_SQL_merge_tmp[, c(StrColumnForHescodes,StrColumnForHescodes)], function(col) grep(paste(sep="","^",VctCodes, collapse='|'), col, ignore.case=FALSE),mc.cores =detectCores()/2 ) ) ## PARALLEL of the above.
} else {
grepoper <-unlist( mclapply( dfmaster_SQL_merge_tmp[, c(StrColumnForHescodes,StrColumnForHescodes)], function(col) grep(paste(sep="","^",VctCodes,"$", collapse='|'), col, ignore.case=FALSE),mc.cores =detectCores()/2 ) ) ## PARALLEL of the above.
}
if(length(grepoper) ==0) {print(" nothing here");return(0) }
dfmaster_SQL_merge_oper <-dfmaster_SQL_merge_tmp[ unique(c(grepoper)),]
## filter Episode Durations on a minimum duration beforehand
if (epidurfilter>0){
dfmaster_SQL_merge_oper <- dfmaster_SQL_merge_oper[abs(round(difftime( dfmaster_SQL_merge_oper$epiend_1, dfmaster_SQL_merge_oper$event_date ,units="days") )) >= epidurfilter,]
}
# remove duplicates
dfmaster_SQL_merge_oper <-dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% filter (! duplicated(event_date)) ### remove duplicate dates.
## History vars:
dfmaster_SQL_merge_oper$HX <-ifelse(dfmaster_SQL_merge_oper$visit_date >= dfmaster_SQL_merge_oper$event_date, 1, 0) ### SET HISTORY VARIABLE (0/1)
dfmaster_SQL_merge_oper <-dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(HXn =sum(HX)) ### Count number of history events
dfmaster_SQL_merge_oper <-dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(HX = max(HX) ) ### replace HX for every duplicating n_eid
dfmaster_SQL_merge_oper$HXd <-ifelse(dfmaster_SQL_merge_oper$visit_date >= dfmaster_SQL_merge_oper$event_date, difftime(dfmaster_SQL_merge_oper$event_date, dfmaster_SQL_merge_oper$visit_date ,units="days") , NA) ### SET FUP time (days) - first event in history
dfmaster_SQL_merge_oper <-suppressWarnings( dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(HXd = -max(abs(HXd), na.rm=T) ) ) ### replace HXd
dfmaster_SQL_merge_oper$HXd[is.infinite(dfmaster_SQL_merge_oper$HXd)]<-NA
# dfmaster_SQL_merge_oper$HXd1 <-ifelse(dfmaster_SQL_merge_oper$visit_date >= dfmaster_SQL_merge_oper$event_date, difftime(dfmaster_SQL_merge_oper$event_date, dfmaster_SQL_merge_oper$visit_date ,units="days") , NA) ### SET FUP time (days) - last event in history
# dfmaster_SQL_merge_oper <-suppressWarnings( dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(HXd1 = -min(abs(HXd1), na.rm=T) ) ) ### replace HXd
# dfmaster_SQL_merge_oper$HXd1[is.infinite(dfmaster_SQL_merge_oper$HXd1)]<-NA
## Episode Duration.
dfmaster_SQL_merge_oper$HXt <-ifelse(dfmaster_SQL_merge_oper$visit_date >= dfmaster_SQL_merge_oper$event_date, round(difftime( dfmaster_SQL_merge_oper$event_date , dfmaster_SQL_merge_oper$epiend_1 ,units="days")) , NA) ### SET FUP time (days) - first event in history
dfmaster_SQL_merge_oper <-suppressWarnings( dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(HXt = -max(abs(HXt), na.rm=T) ) ) ### replace HXd
dfmaster_SQL_merge_oper$HXt[is.infinite(dfmaster_SQL_merge_oper$HXt)]<-NA
# ## total episode dur (tt): ### NOT FOR GP/SR DATA
dfmaster_SQL_merge_oper$episodedur <- abs(round(difftime( dfmaster_SQL_merge_oper$event_date , dfmaster_SQL_merge_oper$epiend_1 ,units="days")))
dfmaster_SQL_merge_oper <-dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(HXto = sum(ifelse (visit_date >= event_date, episodedur ,0 ) ))
#
## follow up vars; 3170028 1834566
dfmaster_SQL_merge_oper$FU <-ifelse(dfmaster_SQL_merge_oper$visit_date < dfmaster_SQL_merge_oper$event_date, 1, 0) ### SET FOLLOW UP VARIABLE (0/1)
dfmaster_SQL_merge_oper <-dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(FUn =sum(FU) ) ### Count number of followup events
dfmaster_SQL_merge_oper <- suppressWarnings(dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(FU = max(FU) )) ### replace FU for every duplicating n_eid
dfmaster_SQL_merge_oper$FUd <-ifelse(dfmaster_SQL_merge_oper$visit_date < dfmaster_SQL_merge_oper$event_date, difftime(dfmaster_SQL_merge_oper$event_date, dfmaster_SQL_merge_oper$visit_date ,units="days") , NA) ### SET FUP time (days) - first event in future.
dfmaster_SQL_merge_oper <-suppressWarnings( dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(FUd = min(FUd, na.rm=T) ) ) ### replace HXd
dfmaster_SQL_merge_oper$FUd[is.infinite(dfmaster_SQL_merge_oper$FUd)]<-NA
# dfmaster_SQL_merge_oper$FUd1 <-ifelse(dfmaster_SQL_merge_oper$visit_date < dfmaster_SQL_merge_oper$event_date, difftime(dfmaster_SQL_merge_oper$event_date, dfmaster_SQL_merge_oper$visit_date ,units="days") , NA) ### SET FUP time (days) - last event in future.
# dfmaster_SQL_merge_oper <-dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(FUd1 = max(FUd1, na.rm=T) ) ### replace HXd
dfmaster_SQL_merge_oper$FUt <-ifelse(dfmaster_SQL_merge_oper$visit_date < dfmaster_SQL_merge_oper$event_date, round(difftime( dfmaster_SQL_merge_oper$epiend_1, dfmaster_SQL_merge_oper$event_date ,units="days") ) , NA) ### SET FUP time (days) - first event in future.
dfmaster_SQL_merge_oper <-suppressWarnings(dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(FUt = min(FUt, na.rm=T) )) ### replace HXd
# ## total episode dur (tt):
dfmaster_SQL_merge_oper <-dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(FUto = sum(ifelse (visit_date < event_date, episodedur ,0 ) ))
## keep uniq
dfmaster_SQL_merge_oper <- dfmaster_SQL_merge_oper[!duplicated(dfmaster_SQL_merge_oper[,DiagnosisVars]),][,DiagnosisVars] ## FILTER UNIQ + KEEP ONLY COLUMNS NEEDED.
# if(recurrence_is_possible==FALSE){
# #TODO
# remove FU* if HXn>0
# }
print(paste(" #found ",nrow(dfmaster_SQL_merge_oper) ))
### ADD DESCRIPTION // RENAME VARIABLES
attr(dfmaster_SQL_merge_oper, "var.labels") <- c("Identifier",paste(StrDescription, colnames(dfmaster_SQL_merge_oper[,-1]), sep = " - ") )
attr(dfmaster_SQL_merge_oper, "var.labels")<-substr(attr(dfmaster_SQL_merge_oper, "var.labels"),0,80)
colnames(dfmaster_SQL_merge_oper)[-1] <- paste(StrTrait, colnames(dfmaster_SQL_merge_oper[,-1]), sep = "_")
save.dta13(dfmaster_SQL_merge_oper, StataOutputFile,compress = TRUE)
return(dfmaster_SQL_merge_oper)
}
#' @export
Death_Masterset<-function(dfmaster_TSDEATHMEDICD10_visitdtonly,StrTrait,StrDescription,StrName,VctCodes,StrFieldcode,StrFieldDateColumn,visitdt,StataOutputFile){
### CHECK IF FILE EXISTS:
if( file.exists(StataOutputFile )) {
print(paste(StataOutputFile," ... already exists!, skipping"))
return(0)
}
print(paste("Death_Masterset - ", StrTrait,":",StrDescription,":",StrFieldcode))
StrColumnNames<-names(dfmaster_TSDEATHMEDICD10_visitdtonly[ grepl(StrFieldcode, names(dfmaster_TSDEATHMEDICD10_visitdtonly))])
#grepoper=c()
#for (i in StrColumnNames )
# grepoper<-c(grepoper,grep(paste(VctCodes, collapse='|'), dfmaster_TSDEATHMEDICD10_visitdtonly[,i], ignore.case=FALSE))
grepoper <-unlist( mclapply( dfmaster_TSDEATHMEDICD10_visitdtonly[, StrColumnNames], function(col) grep(paste(VctCodes, collapse='|'), col, ignore.case=FALSE),mc.cores =detectCores()/2 ) ) ### PARALLEL., the abovve is not parallelized
dfmaster_TSDEATHMEDICD10_visitdtonly_subset<-dfmaster_TSDEATHMEDICD10_visitdtonly[ unique(c(grepoper)),][c("n_eid",visitdt,StrFieldDateColumn)]
if (is.data.frame(dfmaster_TSDEATHMEDICD10_visitdtonly_subset) && nrow(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)>0){
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$grep<-1
### Calculate days to death_event
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$DthFud <-as.vector(difftime(dfmaster_TSDEATHMEDICD10_visitdtonly_subset[,StrFieldDateColumn], dfmaster_TSDEATHMEDICD10_visitdtonly_subset[,visitdt] ,units="days"))
#### REMOVING IF DAYS ARE <0 !!!! TODO FIGURE OUT WHY !!!!
#n_eid with negative death:
#1118475
#1730573 <-- ts_40000_1_0 contains a later date thatn ts_40000_0_0 ?? ( which is the only case.)
#1507226
dfmaster_TSDEATHMEDICD10_visitdtonly_subset<-dfmaster_TSDEATHMEDICD10_visitdtonly_subset[dfmaster_TSDEATHMEDICD10_visitdtonly_subset$DthFud>=0,c(1,4,5)] ## keep only n_eid, death=1 and days until death
### ADD DESCRIPTION // RENAME VARIABLES
names(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)<-c("n_eid",paste(StrTrait,StrName,"FUn",sep="_"),paste(StrTrait,StrName,"FUd",sep="_"))
attr(dfmaster_TSDEATHMEDICD10_visitdtonly_subset, "var.labels") <- c("Identifier",StrDescription,paste(StrDescription,"- FU, Days"))
attr(dfmaster_TSDEATHMEDICD10_visitdtonly_subset, "var.labels")<-substr(attr(dfmaster_TSDEATHMEDICD10_visitdtonly_subset, "var.labels"),0,80)
### filter on NA so when death!=1 (FUd can be missing as visit date for visit 2 is not always available )
head(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)
dfmaster_TSDEATHMEDICD10_visitdtonly_subset<-dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ !is.na(dfmaster_TSDEATHMEDICD10_visitdtonly_subset[,2]),]
head(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)
if(nrow(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)==0){
print("ERROR: could not find any variables")
return(0)
}
save.dta13(dfmaster_TSDEATHMEDICD10_visitdtonly_subset, StataOutputFile,compress = TRUE)
return(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)
} else {
print("ERROR: could not find any variables");
}
}
#' @export
Query_Masterset2<-function(dfmaster_TSDEATHMEDICD10_visitdtonly,StrTrait,StrDescription,StrName,VctTSconditions,StataOutputFile,visitreference,Nvisits){
### CHECK IF FILE EXISTS:
if( file.exists(StataOutputFile )) {
print(paste(StataOutputFile," ... already exists!, skipping"))
return(0)
}
print(paste(StrTrait,":",StrDescription))
VctVisits=0:(Nvisits-1)
LstColumnnamesForVisitDates=paste("ts_53_",VctVisits,"_0",sep="")
### Loop over visits.
df=dfmaster_TSDEATHMEDICD10_visitdtonly[c("n_eid",LstColumnnamesForVisitDates)]
for (visit in VctVisits){
visitcol=paste("_",visit,"_",sep="")
print(visitcol)
### FIND MATCHES, store them in grepoper=c():
grepoper=c()
#### ======= argument ; can be _n3434_=1|2||3
for (Condition in VctTSconditions[grep("=",VctTSconditions)]) {
# Condition=VctTSconditions[grep("=",VctTSconditions)][1]
print( unlist(strsplit(Condition,split="=")) )
StrTScolumn=unlist(strsplit(Condition,split="="))[1]
StrTScodes=unlist(strsplit(Condition,split="="))[2]
StrColumnNames<-names(dfmaster_TSDEATHMEDICD10_visitdtonly[ grepl(StrTScolumn, names(dfmaster_TSDEATHMEDICD10_visitdtonly)) ])
StrColumnNames_pervisit<-StrColumnNames[grepl(visitcol,StrColumnNames)]
if ( length(StrColumnNames) == 0) { stop(paste("ERROR: at least one column missing, there may be more missing, please check!!!: ",paste(StrTScolumn,collapse=",") ) ) }
for (i in StrColumnNames_pervisit ) {
# i = StrColumnNames_pervisit[1]
#for (code in unlist(strsplit(StrTScodes,split="\\|" )) ){
# grepoper<- c(grepoper,which(as.numeric(dfmaster_TSDEATHMEDICD10_visitdtonly[,i])==code))
#}
print(paste(" #visit=",visit, "; #Condition=",Condition, "; #i=",i, "; #found ",sum(as.numeric(dfmaster_TSDEATHMEDICD10_visitdtonly[,i]) %in% unlist(strsplit(StrTScodes,split="\\|" )))) )
grepoper<- c(grepoper,which(as.numeric(dfmaster_TSDEATHMEDICD10_visitdtonly[,i]) %in% unlist(strsplit(StrTScodes,split="\\|" )) )) ## faster.. but the above is similar to the code below.. should figure out how to merge below with this or something..
}
}
# ≥ argument can be _n3434_≥1 , FIX ≤
for (Condition in VctTSconditions[ grep("≥",VctTSconditions) ]) {
print( unlist(strsplit(Condition,split="≥")) )
StrTScolumn=unlist(strsplit(Condition,split="≥"))[1]
StrTScodes=unlist(strsplit(Condition,split="≥"))[2]
StrColumnNames<-names(dfmaster_TSDEATHMEDICD10_visitdtonly[ grepl(StrTScolumn, names(dfmaster_TSDEATHMEDICD10_visitdtonly))])
StrColumnNames_pervisit<-StrColumnNames[grepl(visitcol,StrColumnNames)]
if ( length(StrColumnNames) == 0) { stop(paste("ERROR: at least one column missing, there may be more missing, please check!!!: ",paste(StrTScolumn,collapse=",") ) ) }
for (i in StrColumnNames_pervisit ) {
grepoper<- c(grepoper,which(as.numeric(dfmaster_TSDEATHMEDICD10_visitdtonly[,i])>=StrTScodes))
}
}
#### >>>>> = argument ; can be _n3434_ > 1 ##### FIX <
for (Condition in VctTSconditions[grep(">",VctTSconditions)]) {
print( unlist(strsplit(Condition,split=">")) )
StrTScolumn=unlist(strsplit(Condition,split=">"))[1]
StrTScodes=unlist(strsplit(Condition,split=">"))[2]
StrColumnNames<-names(dfmaster_TSDEATHMEDICD10_visitdtonly[ grepl(StrTScolumn, names(dfmaster_TSDEATHMEDICD10_visitdtonly))])
StrColumnNames_pervisit<-StrColumnNames[grepl(visitcol,StrColumnNames)]
if ( length(StrColumnNames) == 0) { stop(paste("ERROR: at least one column missing, there may be more missing, please check!!!: ",paste(StrTScolumn,collapse=",") ) ) }
for (i in StrColumnNames_pervisit ) {
grepoper<- c(grepoper,which(as.numeric(dfmaster_TSDEATHMEDICD10_visitdtonly[,i])>StrTScodes))
}
}
##########################################
### / put it in dataframe:
if (!length(grepoper)==0){
df$grep<-NA
df[unique(c(grepoper)),]$grep<-1
names(df)[names(df)=="grep"]<-paste(StrName,"_",visit,sep="")
} else {
df$grep<-NA
names(df)[names(df)=="grep"]<-paste(StrName,"_",visit,sep="")
}
}
df<-df[ !apply(df[(ncol(df)-Nvisits+1):ncol(df)], 1, function(x) all(is.na(x))),] ### filter on trait==1 ( df[(ncol(df)-3+1):ncol(df)] filters on TS_*_*. should substitute 3 by #visits)
df$HXn<-0
df$FUn<-0
df$HXd<-NA
df$FUd<-NA
# loop over VctVisits 0 1 2
for (visitnr in VctVisits){ ### replace NA with 0 for the TS_0/1/2 column if they attended the visit. ; leave NA if there is no visit
Coltrait=paste(StrName,visitnr,sep="_")
Colvisit=paste("ts_53_",visitnr,"_0",sep="")
Colreference=paste("ts_53_",visitreference,"_0",sep="") ## TODO: ADD IN OPTION TO USE CUSTOMIZED VECTOR WITH DATES
df[ is.na(df[Coltrait]) & !is.na(df[Colvisit]) ,Coltrait]<-0
if (visitnr<=visitreference ){ ## TODO: ADD IN OPTION TO USE CUSTOMIZED VECTOR WITH DATES -- replace this loop with something more generic?
HX_TF<- df[Coltrait]==1 & !is.na(df[Coltrait])
df[ HX_TF , ]$HXn<-1+df[ HX_TF, ]$HXn
HXd_TF<-!is.na(df[c(Colvisit)]) & !is.na(df[c(Colreference)]) & df[Colvisit] <= df[Colreference] & df[Coltrait]==1
df[HXd_TF & is.na(df$HXd), ]$HXd <- df[HXd_TF & is.na(df$HXd), Colvisit]-df[HXd_TF & is.na(df$HXd), Colreference] ### skip if HXd is filled in, to keep first observation.
}
if (visitnr>visitreference ){
FU_TF<- df[Coltrait]==1 & !is.na(df[Coltrait])
df[FU_TF, ]$FUn<- 1+ df[ FU_TF, ]$FUn
FUd_TF<- !is.na(df[c(Colvisit)]) & !is.na(df[c(Colreference)]) & df[Colvisit] > df[Colreference] & df[Coltrait]==1
df[FUd_TF & is.na(df$FUd), ]$FUd<- df[FUd_TF & is.na(df$FUd), Colvisit]-df[FUd_TF & is.na(df$FUd), Colreference] ### skip if FUd is filled in, to keep first observation.
}
}
df$FUdlv<-NA ## calculate days to last visit before FUd and discrepancies between Q's
df$HXdlv<-NA ## calculate days to last visit before FUd and discrepancies between Q's
df$ERR<-NA ## detect individuals with questionable answers (010, 100,101 etc.)
df$ERRtmp<-df[,paste(StrName,"0",sep="_")]
for (visitnr in VctVisits){ ### replace NA with 0 for the TS_0/1/2 column if they attended the visit. ; leave NA if there is no visit
Coltrait=paste(StrName,visitnr,sep="_")
Colvisit=paste("ts_53_",visitnr,"_0",sep="")
Colreference=paste("ts_53_",visitreference,"_0",sep="")
Colvisitreferencediff<-!is.na(df[c(Colvisit)]) & !is.na(df[c(Colreference)]) & !is.na(df$FUd) & (df[,Colvisit]-df[,Colreference]) < df$FUd
df[Colvisitreferencediff,]$FUdlv<-df[Colvisitreferencediff,Colvisit ] - df[Colvisitreferencediff,Colreference ]
Colvisitreferencediff<-!is.na(df[c(Colvisit)]) & !is.na(df[c(Colreference)]) & !is.na(df$HXd) & (df[,Colvisit]-df[,Colreference]) < df$HXd
df[Colvisitreferencediff,]$HXdlv<-df[Colvisitreferencediff,Colvisit ] - df[Colvisitreferencediff,Colreference ]
df[ !is.na(df[c(Colvisit)]) & df$ERRtmp > df[,Coltrait] ,"ERR"]<-1
df[ !is.na(df[c(Colvisit)]) ,"ERRtmp"]<-df[ !is.na(df[c(Colvisit)]) ,Coltrait]
}
if( is.data.frame(df) && nrow(df)>0){
DiagnosisVars=c("n_eid",paste(StrName,visitreference,sep="_"),"HXn","HXd","HXdlv","FUn","FUd","FUdlv","ERR") ### variables you wil create and you want to keep later...
DiagnosisVarsDescription=c("Identifier","Yes/No on VisitDate","History cases answered yes","History days before first observation","History to RefDate in days", "FollowUp cases answered yes","FollowUp from RefDate in days","FollowUp; last followup before first occurence in days","Erroneous individuals, questionable answers")
df<-df[,DiagnosisVars] ## KEEP ONLY COLUMNS NEEDED.
names(df)[-1]<-c(paste(StrTrait,"_",StrName,sep="") ,paste(StrTrait,"_",StrName,"_",names(df[-1:-2]),sep=""))
### ADD DESCRIPTION // RENAME VARIABLES
# attr(df, "var.labels") <- c("Identifier",paste(StrDescription,DiagnosisVarsDescription[-1],colapse="",sep=" "))
attr(df, "var.labels") <- c("Identifier",paste(StrDescription,DiagnosisVarsDescription[-1],colapse="",sep=" "))
attr(df, "var.labels")<-substr(attr(df, "var.labels"),0,80)
save.dta13(df, StataOutputFile,compress = TRUE)
return(df)
} else {
stop(paste("ERROR: could not find any variables"),paste(VctTSconditions,collapse=","))
}
}
#' @export
CheckIfColumnsAreAvailable<-function(df,VctColumnames){
#TODO check if columns exists before running functions!!
return(0)
}
#' @export
Query_TSAmin_masterset<-function(dfmaster_TSDEATHMEDICD10_visitdtonly,StrTrait,StrDescription,StrName,VctTSAcolumns,visitdt,StataOutputFile){
#dfmaster_TSDEATHMEDICD10_visitdtonly<-UKbioDatasetm
#StrTrait<-"a1def"
#StrDescription<-"asdsad"
#VctTSAcolumns<-"n_22152_"
#visitdt<-paste("ts_53_",2,"_0",sep="")
#StataOutputFile<-"/data_work/databases/ukbiobanks/dataset/"
### CHECK IF FILE EXISTS:
if( file.exists(StataOutputFile )) {
print(paste(StataOutputFile," ... already exists!, skipping"))
return(0)
}
print(paste(StrTrait,":",StrDescription))
dfmaster_TSDEATHMEDICD10_visitdtonly_subset<-dfmaster_TSDEATHMEDICD10_visitdtonly[ , grepl( paste(c("n_eid",visitdt,VctTSAcolumns), collapse='|') , names( dfmaster_TSDEATHMEDICD10_visitdtonly ) ) ]
### BirthDate:
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$Birthdate<- as.Date(paste('01', paste(dfmaster_TSDEATHMEDICD10_visitdtonly$n_52_0_0,dfmaster_TSDEATHMEDICD10_visitdtonly$n_34_0_0)), format='%d %m %Y')
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$Age<-(dfmaster_TSDEATHMEDICD10_visitdtonly_subset[,visitdt] - dfmaster_TSDEATHMEDICD10_visitdtonly_subset$Birthdate) /365.25
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAOveralMin<-NA ## set initial value
if( ncol(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)==5) {stop("Missing TSA column? cannot find anything; so at least 1 missing")} ## 0 column ; error; missing column.
for (VctTSAcolumn in VctTSAcolumns){
print(paste("mean-minimum age across columns:",VctTSAcolumn))
TSAcolumns_TF=grepl( paste(c(VctTSAcolumn), collapse='|') , names( dfmaster_TSDEATHMEDICD10_visitdtonly_subset ) )
if(length(which(TSAcolumns_TF))==1){ ## if only one visit is available no mean is possible; then duplicate it.
print("Only one visit is available.")
dfmaster_TSDEATHMEDICD10_visitdtonly_subset[paste(VctTSAcolumn,"tmp_0",sep="")]<-dfmaster_TSDEATHMEDICD10_visitdtonly_subset[TSAcolumns_TF]
TSAcolumns_TF=grepl( paste(c(VctTSAcolumn), collapse='|') , names( dfmaster_TSDEATHMEDICD10_visitdtonly_subset ) )
}
dfmaster_TSDEATHMEDICD10_visitdtonly_subset[TSAcolumns_TF][dfmaster_TSDEATHMEDICD10_visitdtonly_subset[TSAcolumns_TF] == -1]<-NA
dfmaster_TSDEATHMEDICD10_visitdtonly_subset[TSAcolumns_TF][dfmaster_TSDEATHMEDICD10_visitdtonly_subset[TSAcolumns_TF] == -3]<-NA
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAmin<-rowMins(as.matrix( dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ ,grepl(paste(VctTSAcolumn, collapse='|'),names(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)) ]),na.rm=TRUE)
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAmax<-rowMaxs(as.matrix( dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ ,grepl(paste(VctTSAcolumn, collapse='|'),names(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)) ]),na.rm=TRUE)
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSADiff<-dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAmax-dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAmin
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAmean<-rowMeans(as.matrix( dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ ,grepl(paste(VctTSAcolumn, collapse='|'),names(dfmaster_TSDEATHMEDICD10_visitdtonly_subset))]),na.rm=TRUE)
### filter outliers: difference between visits >5years.
outliers<-dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSADiff>5 & !is.infinite(dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSADiff)
if(length(which(outliers))>0){## check if outliers exist.
print("removing TSAd outliers based on age-difference>5")
dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ outliers,]$TSAmean<-NA
}
#dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ is.infinite(dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAmean),]$TSAmean<-NA
### add minimum to existing column;
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAOveralMin<-rowMins(as.matrix( cbind(dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAmean,dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAOveralMin)),na.rm=TRUE)
dfmaster_TSDEATHMEDICD10_visitdtonly_subset[is.infinite(dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAOveralMin),]$TSAOveralMin<-NA
}
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd<-as.numeric(round((dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAOveralMin - dfmaster_TSDEATHMEDICD10_visitdtonly_subset$Age)*365.25))
# dfmaster_TSDEATHMEDICD10_visitdtonly_subset<-dfmaster_TSDEATHMEDICD10_visitdtonly_subset[dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAmean>=0,] # filter age>0?
dfmaster_TSDEATHMEDICD10_visitdtonly_subset<-(dfmaster_TSDEATHMEDICD10_visitdtonly_subset[!is.na(dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd),c("n_eid","TSAd")])
### Recalculate to Days from Reference.
dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ abs(dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd)<365.25 & dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd>0, "TSAd"] <- dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ abs(dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd)<365.25 & dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd>0, "TSAd"] /2 # within 1 year of diagnosis
dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ abs(dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd)<365.25 & dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd<0, "TSAd"] <- dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ abs(dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd)<365.25 & dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd<0, "TSAd"] /2 # within 1 year of diagnosis
dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd <= -365.25 ,"TSAd"]<-dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd <= -365.25 ,"TSAd"]+round(365.25/2) ## beyond one year in history
dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd >= 365.25 ,"TSAd"]<-dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd >= 365.25 ,"TSAd"]-round(365.25/2) ## beyond one year in future
if( is.data.frame(dfmaster_TSDEATHMEDICD10_visitdtonly_subset) && nrow(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)>0){
names(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)<-c("n_eid",paste(StrTrait,StrName,sep="_"))
attr(dfmaster_TSDEATHMEDICD10_visitdtonly_subset, "var.labels") <- c("Identifier",StrDescription)
attr(dfmaster_TSDEATHMEDICD10_visitdtonly_subset, "var.labels")<-substr(attr(dfmaster_TSDEATHMEDICD10_visitdtonly_subset, "var.labels"),0,80)
save.dta13(dfmaster_TSDEATHMEDICD10_visitdtonly_subset, StataOutputFile,compress = TRUE)
return(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)
} else {
print(paste("WARNING: df empty; could not find any variables OR observations (could be due to visit 2 having no observations vs visit 1), maybe check: ",StrTrait,"///",paste(VctTSAcolumn,collapse=",")))
}
}
#' #' Extracting Vars From larger UK Bio dataset
#' #'
#' #' Extracts vars from ukbiobank from larger set of variables based on selection.
#' #'
#' #' @param UKbioDataset df
#' #' @param dfDefinitions_VarSelection df
#' #' @param StataOutputFile file
#' #' @keywords ExtractVarsFromMasterSet CreateUKBiobankPhentoypes
#' #' @return None
#' #'
#' #' @examples
#' #' ExtractVarsFromMasterSet(UKbioDataset,dfDefinitions_VarSelection,StataOutputFile)
#' #'
#' #' @export
#' ExtractVarsFromMasterSet<-function(UKbioDataset,dfDefinitions_VarSelection,StataOutputFile){
#'
#' #OutputdirMerged=paste(Outputdir,"/merged/",sep="")
#' #StataOutputFile=paste(OutputdirMerged,"/DataSelection.dta",sep="")
#'
#' ############################################################
#' ############ EXTRACT SELECTION FROM MASTERSET:
#' dfSQL_TraitTable_VarSelection_Selected<-dfDefinitions_VarSelection[dfDefinitions_VarSelection$Include==1,]
#' #as.vector(dfSQL_TraitTable_VarSelection_Selected$Var)
#' ### FOUND:
#' KeyAvail=names(UKbioDataset) %in% as.vector(dfSQL_TraitTable_VarSelection_Selected$var)
#' Avail=(names(UKbioDataset)[KeyAvail])
#' dfmasterSelection<-UKbioDataset[KeyAvail]
#' attr(dfmasterSelection,"var.labels")<-attr(UKbioDataset,"var.labels")[KeyAvail]
#' attr(dfmasterSelection, "var.labels")<-substr(attr(dfmasterSelection, "var.labels"),0,80)
#'
#' ### NOT FOUND
#' NotAvail=as.vector(dfSQL_TraitTable_VarSelection_Selected$Var[!as.vector(dfSQL_TraitTable_VarSelection_Selected$var) %in% names(UKbioDataset) ])
#'
#' print(paste("FOUND:",paste(Avail,collapse=", ")))
#' print(paste("MISSING:",paste(NotAvail,collapse=", ")))
#'
#' save.dta13(dfmasterSelection,StataOutputFile,compress = TRUE)
#'
#' }
niekverw/ukpheno documentation built on May 4, 2020, 9:01 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions Query_TSAmin_masterset CheckIfColumnsAreAvailable Query_Masterset2 Death_Masterset Outcome_HES
#' @export
Outcome_HES<-function(dfmaster_SQL_merge,StrTrait,StrDescription,VctCodes,epidurfilter=0,StrColumnForHescodes,StrColumnnameForEventDate,StrColumnnameForVisitDate,StataOutputFile,
grep_with_wildcard=TRUE #recurrence_is_possible=TRUE # not for self reported data..
){
# epidurfilter=0
# dfmaster_SQL_merge=dfgpclinical
# dfmaster_SQL_merge=dfhesintables
### CHECK IF FILE EXISTS:
if( file.exists(StataOutputFile )) {
print(paste(StataOutputFile," ... already exists!, skipping"))
return(0)
}
print(paste("extracting outcomes from episode data, checking codes:",StrTrait,":",StrDescription,"-",StrColumnnameForVisitDate, "--",paste(StrColumnForHescodes,collapse=",") ) )
DiagnosisVars=c("n_eid","HXn","HXd","HXt","HXto","FUn","FUd","FUt","FUto") ### variables you wil create and you want to keep later...
### FUt/ FUto = NOT FOR GP DATA, only HESIN
#### RENAME COLUMNS
names(dfmaster_SQL_merge)[names(dfmaster_SQL_merge)==StrColumnnameForEventDate]<-"event_date"
names(dfmaster_SQL_merge)[names(dfmaster_SQL_merge)==StrColumnnameForVisitDate]<-"visit_date"
## remove records that have no visit date (relevant for ts_53_1_0 , ts_53_2_0).
dfmaster_SQL_merge_tmp<-dfmaster_SQL_merge[!is.na(dfmaster_SQL_merge$visit_date),]
########### GREP CODE AT THE COLUMNS 'StrColumnForHescodes' ;; TODO grep should be ^: paste(sep="","^",VctCodes, collapse='|')
#grepoper=c()
#for (i in StrColumnForHescodes ){
# grepoper<-c(grepoper,grep(paste(VctCodes, collapse='|'), dfmaster_SQL_merge_tmp[,i], ignore.case=FALSE))
#}
if (length(StrColumnForHescodes)==1){StrColumnForHescodes = c(StrColumnForHescodes,StrColumnForHescodes)}
if(grep_with_wildcard==TRUE){
grepoper <-unlist( mclapply( dfmaster_SQL_merge_tmp[, c(StrColumnForHescodes,StrColumnForHescodes)], function(col) grep(paste(sep="","^",VctCodes, collapse='|'), col, ignore.case=FALSE),mc.cores =detectCores()/2 ) ) ## PARALLEL of the above.
} else {
grepoper <-unlist( mclapply( dfmaster_SQL_merge_tmp[, c(StrColumnForHescodes,StrColumnForHescodes)], function(col) grep(paste(sep="","^",VctCodes,"$", collapse='|'), col, ignore.case=FALSE),mc.cores =detectCores()/2 ) ) ## PARALLEL of the above.
}
if(length(grepoper) ==0) {print(" nothing here");return(0) }
dfmaster_SQL_merge_oper <-dfmaster_SQL_merge_tmp[ unique(c(grepoper)),]
## filter Episode Durations on a minimum duration beforehand
if (epidurfilter>0){
dfmaster_SQL_merge_oper <- dfmaster_SQL_merge_oper[abs(round(difftime( dfmaster_SQL_merge_oper$epiend_1, dfmaster_SQL_merge_oper$event_date ,units="days") )) >= epidurfilter,]
}
# remove duplicates
dfmaster_SQL_merge_oper <-dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% filter (! duplicated(event_date)) ### remove duplicate dates.
## History vars:
dfmaster_SQL_merge_oper$HX <-ifelse(dfmaster_SQL_merge_oper$visit_date >= dfmaster_SQL_merge_oper$event_date, 1, 0) ### SET HISTORY VARIABLE (0/1)
dfmaster_SQL_merge_oper <-dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(HXn =sum(HX)) ### Count number of history events
dfmaster_SQL_merge_oper <-dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(HX = max(HX) ) ### replace HX for every duplicating n_eid
dfmaster_SQL_merge_oper$HXd <-ifelse(dfmaster_SQL_merge_oper$visit_date >= dfmaster_SQL_merge_oper$event_date, difftime(dfmaster_SQL_merge_oper$event_date, dfmaster_SQL_merge_oper$visit_date ,units="days") , NA) ### SET FUP time (days) - first event in history
dfmaster_SQL_merge_oper <-suppressWarnings( dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(HXd = -max(abs(HXd), na.rm=T) ) ) ### replace HXd
dfmaster_SQL_merge_oper$HXd[is.infinite(dfmaster_SQL_merge_oper$HXd)]<-NA
# dfmaster_SQL_merge_oper$HXd1 <-ifelse(dfmaster_SQL_merge_oper$visit_date >= dfmaster_SQL_merge_oper$event_date, difftime(dfmaster_SQL_merge_oper$event_date, dfmaster_SQL_merge_oper$visit_date ,units="days") , NA) ### SET FUP time (days) - last event in history
# dfmaster_SQL_merge_oper <-suppressWarnings( dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(HXd1 = -min(abs(HXd1), na.rm=T) ) ) ### replace HXd
# dfmaster_SQL_merge_oper$HXd1[is.infinite(dfmaster_SQL_merge_oper$HXd1)]<-NA
## Episode Duration.
dfmaster_SQL_merge_oper$HXt <-ifelse(dfmaster_SQL_merge_oper$visit_date >= dfmaster_SQL_merge_oper$event_date, round(difftime( dfmaster_SQL_merge_oper$event_date , dfmaster_SQL_merge_oper$epiend_1 ,units="days")) , NA) ### SET FUP time (days) - first event in history
dfmaster_SQL_merge_oper <-suppressWarnings( dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(HXt = -max(abs(HXt), na.rm=T) ) ) ### replace HXd
dfmaster_SQL_merge_oper$HXt[is.infinite(dfmaster_SQL_merge_oper$HXt)]<-NA
# ## total episode dur (tt): ### NOT FOR GP/SR DATA
dfmaster_SQL_merge_oper$episodedur <- abs(round(difftime( dfmaster_SQL_merge_oper$event_date , dfmaster_SQL_merge_oper$epiend_1 ,units="days")))
dfmaster_SQL_merge_oper <-dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(HXto = sum(ifelse (visit_date >= event_date, episodedur ,0 ) ))
#
## follow up vars; 3170028 1834566
dfmaster_SQL_merge_oper$FU <-ifelse(dfmaster_SQL_merge_oper$visit_date < dfmaster_SQL_merge_oper$event_date, 1, 0) ### SET FOLLOW UP VARIABLE (0/1)
dfmaster_SQL_merge_oper <-dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(FUn =sum(FU) ) ### Count number of followup events
dfmaster_SQL_merge_oper <- suppressWarnings(dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(FU = max(FU) )) ### replace FU for every duplicating n_eid
dfmaster_SQL_merge_oper$FUd <-ifelse(dfmaster_SQL_merge_oper$visit_date < dfmaster_SQL_merge_oper$event_date, difftime(dfmaster_SQL_merge_oper$event_date, dfmaster_SQL_merge_oper$visit_date ,units="days") , NA) ### SET FUP time (days) - first event in future.
dfmaster_SQL_merge_oper <-suppressWarnings( dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(FUd = min(FUd, na.rm=T) ) ) ### replace HXd
dfmaster_SQL_merge_oper$FUd[is.infinite(dfmaster_SQL_merge_oper$FUd)]<-NA
# dfmaster_SQL_merge_oper$FUd1 <-ifelse(dfmaster_SQL_merge_oper$visit_date < dfmaster_SQL_merge_oper$event_date, difftime(dfmaster_SQL_merge_oper$event_date, dfmaster_SQL_merge_oper$visit_date ,units="days") , NA) ### SET FUP time (days) - last event in future.
# dfmaster_SQL_merge_oper <-dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(FUd1 = max(FUd1, na.rm=T) ) ### replace HXd
dfmaster_SQL_merge_oper$FUt <-ifelse(dfmaster_SQL_merge_oper$visit_date < dfmaster_SQL_merge_oper$event_date, round(difftime( dfmaster_SQL_merge_oper$epiend_1, dfmaster_SQL_merge_oper$event_date ,units="days") ) , NA) ### SET FUP time (days) - first event in future.
dfmaster_SQL_merge_oper <-suppressWarnings(dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(FUt = min(FUt, na.rm=T) )) ### replace HXd
# ## total episode dur (tt):
dfmaster_SQL_merge_oper <-dfmaster_SQL_merge_oper %>% group_by(n_eid) %>% mutate(FUto = sum(ifelse (visit_date < event_date, episodedur ,0 ) ))
## keep uniq
dfmaster_SQL_merge_oper <- dfmaster_SQL_merge_oper[!duplicated(dfmaster_SQL_merge_oper[,DiagnosisVars]),][,DiagnosisVars] ## FILTER UNIQ + KEEP ONLY COLUMNS NEEDED.
# if(recurrence_is_possible==FALSE){
# #TODO
# remove FU* if HXn>0
# }
print(paste(" #found ",nrow(dfmaster_SQL_merge_oper) ))
### ADD DESCRIPTION // RENAME VARIABLES
attr(dfmaster_SQL_merge_oper, "var.labels") <- c("Identifier",paste(StrDescription, colnames(dfmaster_SQL_merge_oper[,-1]), sep = " - ") )
attr(dfmaster_SQL_merge_oper, "var.labels")<-substr(attr(dfmaster_SQL_merge_oper, "var.labels"),0,80)
colnames(dfmaster_SQL_merge_oper)[-1] <- paste(StrTrait, colnames(dfmaster_SQL_merge_oper[,-1]), sep = "_")
save.dta13(dfmaster_SQL_merge_oper, StataOutputFile,compress = TRUE)
return(dfmaster_SQL_merge_oper)
}
#' @export
Death_Masterset<-function(dfmaster_TSDEATHMEDICD10_visitdtonly,StrTrait,StrDescription,StrName,VctCodes,StrFieldcode,StrFieldDateColumn,visitdt,StataOutputFile){
### CHECK IF FILE EXISTS:
if( file.exists(StataOutputFile )) {
print(paste(StataOutputFile," ... already exists!, skipping"))
return(0)
}
print(paste("Death_Masterset - ", StrTrait,":",StrDescription,":",StrFieldcode))
StrColumnNames<-names(dfmaster_TSDEATHMEDICD10_visitdtonly[ grepl(StrFieldcode, names(dfmaster_TSDEATHMEDICD10_visitdtonly))])
#grepoper=c()
#for (i in StrColumnNames )
# grepoper<-c(grepoper,grep(paste(VctCodes, collapse='|'), dfmaster_TSDEATHMEDICD10_visitdtonly[,i], ignore.case=FALSE))
grepoper <-unlist( mclapply( dfmaster_TSDEATHMEDICD10_visitdtonly[, StrColumnNames], function(col) grep(paste(VctCodes, collapse='|'), col, ignore.case=FALSE),mc.cores =detectCores()/2 ) ) ### PARALLEL., the abovve is not parallelized
dfmaster_TSDEATHMEDICD10_visitdtonly_subset<-dfmaster_TSDEATHMEDICD10_visitdtonly[ unique(c(grepoper)),][c("n_eid",visitdt,StrFieldDateColumn)]
if (is.data.frame(dfmaster_TSDEATHMEDICD10_visitdtonly_subset) && nrow(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)>0){
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$grep<-1
### Calculate days to death_event
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$DthFud <-as.vector(difftime(dfmaster_TSDEATHMEDICD10_visitdtonly_subset[,StrFieldDateColumn], dfmaster_TSDEATHMEDICD10_visitdtonly_subset[,visitdt] ,units="days"))
#### REMOVING IF DAYS ARE <0 !!!! TODO FIGURE OUT WHY !!!!
#n_eid with negative death:
#1118475
#1730573 <-- ts_40000_1_0 contains a later date thatn ts_40000_0_0 ?? ( which is the only case.)
#1507226
dfmaster_TSDEATHMEDICD10_visitdtonly_subset<-dfmaster_TSDEATHMEDICD10_visitdtonly_subset[dfmaster_TSDEATHMEDICD10_visitdtonly_subset$DthFud>=0,c(1,4,5)] ## keep only n_eid, death=1 and days until death
### ADD DESCRIPTION // RENAME VARIABLES
names(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)<-c("n_eid",paste(StrTrait,StrName,"FUn",sep="_"),paste(StrTrait,StrName,"FUd",sep="_"))
attr(dfmaster_TSDEATHMEDICD10_visitdtonly_subset, "var.labels") <- c("Identifier",StrDescription,paste(StrDescription,"- FU, Days"))
attr(dfmaster_TSDEATHMEDICD10_visitdtonly_subset, "var.labels")<-substr(attr(dfmaster_TSDEATHMEDICD10_visitdtonly_subset, "var.labels"),0,80)
### filter on NA so when death!=1 (FUd can be missing as visit date for visit 2 is not always available )
head(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)
dfmaster_TSDEATHMEDICD10_visitdtonly_subset<-dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ !is.na(dfmaster_TSDEATHMEDICD10_visitdtonly_subset[,2]),]
head(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)
if(nrow(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)==0){
print("ERROR: could not find any variables")
return(0)
}
save.dta13(dfmaster_TSDEATHMEDICD10_visitdtonly_subset, StataOutputFile,compress = TRUE)
return(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)
} else {
print("ERROR: could not find any variables");
}
}
#' @export
Query_Masterset2<-function(dfmaster_TSDEATHMEDICD10_visitdtonly,StrTrait,StrDescription,StrName,VctTSconditions,StataOutputFile,visitreference,Nvisits){
### CHECK IF FILE EXISTS:
if( file.exists(StataOutputFile )) {
print(paste(StataOutputFile," ... already exists!, skipping"))
return(0)
}
print(paste(StrTrait,":",StrDescription))
VctVisits=0:(Nvisits-1)
LstColumnnamesForVisitDates=paste("ts_53_",VctVisits,"_0",sep="")
### Loop over visits.
df=dfmaster_TSDEATHMEDICD10_visitdtonly[c("n_eid",LstColumnnamesForVisitDates)]
for (visit in VctVisits){
visitcol=paste("_",visit,"_",sep="")
print(visitcol)
### FIND MATCHES, store them in grepoper=c():
grepoper=c()
#### ======= argument ; can be _n3434_=1|2||3
for (Condition in VctTSconditions[grep("=",VctTSconditions)]) {
# Condition=VctTSconditions[grep("=",VctTSconditions)][1]
print( unlist(strsplit(Condition,split="=")) )
StrTScolumn=unlist(strsplit(Condition,split="="))[1]
StrTScodes=unlist(strsplit(Condition,split="="))[2]
StrColumnNames<-names(dfmaster_TSDEATHMEDICD10_visitdtonly[ grepl(StrTScolumn, names(dfmaster_TSDEATHMEDICD10_visitdtonly)) ])
StrColumnNames_pervisit<-StrColumnNames[grepl(visitcol,StrColumnNames)]
if ( length(StrColumnNames) == 0) { stop(paste("ERROR: at least one column missing, there may be more missing, please check!!!: ",paste(StrTScolumn,collapse=",") ) ) }
for (i in StrColumnNames_pervisit ) {
# i = StrColumnNames_pervisit[1]
#for (code in unlist(strsplit(StrTScodes,split="\\|" )) ){
# grepoper<- c(grepoper,which(as.numeric(dfmaster_TSDEATHMEDICD10_visitdtonly[,i])==code))
#}
print(paste(" #visit=",visit, "; #Condition=",Condition, "; #i=",i, "; #found ",sum(as.numeric(dfmaster_TSDEATHMEDICD10_visitdtonly[,i]) %in% unlist(strsplit(StrTScodes,split="\\|" )))) )
grepoper<- c(grepoper,which(as.numeric(dfmaster_TSDEATHMEDICD10_visitdtonly[,i]) %in% unlist(strsplit(StrTScodes,split="\\|" )) )) ## faster.. but the above is similar to the code below.. should figure out how to merge below with this or something..
}
}
# ≥ argument can be _n3434_≥1 , FIX ≤
for (Condition in VctTSconditions[ grep("≥",VctTSconditions) ]) {
print( unlist(strsplit(Condition,split="≥")) )
StrTScolumn=unlist(strsplit(Condition,split="≥"))[1]
StrTScodes=unlist(strsplit(Condition,split="≥"))[2]
StrColumnNames<-names(dfmaster_TSDEATHMEDICD10_visitdtonly[ grepl(StrTScolumn, names(dfmaster_TSDEATHMEDICD10_visitdtonly))])
StrColumnNames_pervisit<-StrColumnNames[grepl(visitcol,StrColumnNames)]
if ( length(StrColumnNames) == 0) { stop(paste("ERROR: at least one column missing, there may be more missing, please check!!!: ",paste(StrTScolumn,collapse=",") ) ) }
for (i in StrColumnNames_pervisit ) {
grepoper<- c(grepoper,which(as.numeric(dfmaster_TSDEATHMEDICD10_visitdtonly[,i])>=StrTScodes))
}
}
#### >>>>> = argument ; can be _n3434_ > 1 ##### FIX <
for (Condition in VctTSconditions[grep(">",VctTSconditions)]) {
print( unlist(strsplit(Condition,split=">")) )
StrTScolumn=unlist(strsplit(Condition,split=">"))[1]
StrTScodes=unlist(strsplit(Condition,split=">"))[2]
StrColumnNames<-names(dfmaster_TSDEATHMEDICD10_visitdtonly[ grepl(StrTScolumn, names(dfmaster_TSDEATHMEDICD10_visitdtonly))])
StrColumnNames_pervisit<-StrColumnNames[grepl(visitcol,StrColumnNames)]
if ( length(StrColumnNames) == 0) { stop(paste("ERROR: at least one column missing, there may be more missing, please check!!!: ",paste(StrTScolumn,collapse=",") ) ) }
for (i in StrColumnNames_pervisit ) {
grepoper<- c(grepoper,which(as.numeric(dfmaster_TSDEATHMEDICD10_visitdtonly[,i])>StrTScodes))
}
}
##########################################
### / put it in dataframe:
if (!length(grepoper)==0){
df$grep<-NA
df[unique(c(grepoper)),]$grep<-1
names(df)[names(df)=="grep"]<-paste(StrName,"_",visit,sep="")
} else {
df$grep<-NA
names(df)[names(df)=="grep"]<-paste(StrName,"_",visit,sep="")
}
}
df<-df[ !apply(df[(ncol(df)-Nvisits+1):ncol(df)], 1, function(x) all(is.na(x))),] ### filter on trait==1 ( df[(ncol(df)-3+1):ncol(df)] filters on TS_*_*. should substitute 3 by #visits)
df$HXn<-0
df$FUn<-0
df$HXd<-NA
df$FUd<-NA
# loop over VctVisits 0 1 2
for (visitnr in VctVisits){ ### replace NA with 0 for the TS_0/1/2 column if they attended the visit. ; leave NA if there is no visit
Coltrait=paste(StrName,visitnr,sep="_")
Colvisit=paste("ts_53_",visitnr,"_0",sep="")
Colreference=paste("ts_53_",visitreference,"_0",sep="") ## TODO: ADD IN OPTION TO USE CUSTOMIZED VECTOR WITH DATES
df[ is.na(df[Coltrait]) & !is.na(df[Colvisit]) ,Coltrait]<-0
if (visitnr<=visitreference ){ ## TODO: ADD IN OPTION TO USE CUSTOMIZED VECTOR WITH DATES -- replace this loop with something more generic?
HX_TF<- df[Coltrait]==1 & !is.na(df[Coltrait])
df[ HX_TF , ]$HXn<-1+df[ HX_TF, ]$HXn
HXd_TF<-!is.na(df[c(Colvisit)]) & !is.na(df[c(Colreference)]) & df[Colvisit] <= df[Colreference] & df[Coltrait]==1
df[HXd_TF & is.na(df$HXd), ]$HXd <- df[HXd_TF & is.na(df$HXd), Colvisit]-df[HXd_TF & is.na(df$HXd), Colreference] ### skip if HXd is filled in, to keep first observation.
}
if (visitnr>visitreference ){
FU_TF<- df[Coltrait]==1 & !is.na(df[Coltrait])
df[FU_TF, ]$FUn<- 1+ df[ FU_TF, ]$FUn
FUd_TF<- !is.na(df[c(Colvisit)]) & !is.na(df[c(Colreference)]) & df[Colvisit] > df[Colreference] & df[Coltrait]==1
df[FUd_TF & is.na(df$FUd), ]$FUd<- df[FUd_TF & is.na(df$FUd), Colvisit]-df[FUd_TF & is.na(df$FUd), Colreference] ### skip if FUd is filled in, to keep first observation.
}
}
df$FUdlv<-NA ## calculate days to last visit before FUd and discrepancies between Q's
df$HXdlv<-NA ## calculate days to last visit before FUd and discrepancies between Q's
df$ERR<-NA ## detect individuals with questionable answers (010, 100,101 etc.)
df$ERRtmp<-df[,paste(StrName,"0",sep="_")]
for (visitnr in VctVisits){ ### replace NA with 0 for the TS_0/1/2 column if they attended the visit. ; leave NA if there is no visit
Coltrait=paste(StrName,visitnr,sep="_")
Colvisit=paste("ts_53_",visitnr,"_0",sep="")
Colreference=paste("ts_53_",visitreference,"_0",sep="")
Colvisitreferencediff<-!is.na(df[c(Colvisit)]) & !is.na(df[c(Colreference)]) & !is.na(df$FUd) & (df[,Colvisit]-df[,Colreference]) < df$FUd
df[Colvisitreferencediff,]$FUdlv<-df[Colvisitreferencediff,Colvisit ] - df[Colvisitreferencediff,Colreference ]
Colvisitreferencediff<-!is.na(df[c(Colvisit)]) & !is.na(df[c(Colreference)]) & !is.na(df$HXd) & (df[,Colvisit]-df[,Colreference]) < df$HXd
df[Colvisitreferencediff,]$HXdlv<-df[Colvisitreferencediff,Colvisit ] - df[Colvisitreferencediff,Colreference ]
df[ !is.na(df[c(Colvisit)]) & df$ERRtmp > df[,Coltrait] ,"ERR"]<-1
df[ !is.na(df[c(Colvisit)]) ,"ERRtmp"]<-df[ !is.na(df[c(Colvisit)]) ,Coltrait]
}
if( is.data.frame(df) && nrow(df)>0){
DiagnosisVars=c("n_eid",paste(StrName,visitreference,sep="_"),"HXn","HXd","HXdlv","FUn","FUd","FUdlv","ERR") ### variables you wil create and you want to keep later...
DiagnosisVarsDescription=c("Identifier","Yes/No on VisitDate","History cases answered yes","History days before first observation","History to RefDate in days", "FollowUp cases answered yes","FollowUp from RefDate in days","FollowUp; last followup before first occurence in days","Erroneous individuals, questionable answers")
df<-df[,DiagnosisVars] ## KEEP ONLY COLUMNS NEEDED.
names(df)[-1]<-c(paste(StrTrait,"_",StrName,sep="") ,paste(StrTrait,"_",StrName,"_",names(df[-1:-2]),sep=""))
### ADD DESCRIPTION // RENAME VARIABLES
# attr(df, "var.labels") <- c("Identifier",paste(StrDescription,DiagnosisVarsDescription[-1],colapse="",sep=" "))
attr(df, "var.labels") <- c("Identifier",paste(StrDescription,DiagnosisVarsDescription[-1],colapse="",sep=" "))
attr(df, "var.labels")<-substr(attr(df, "var.labels"),0,80)
save.dta13(df, StataOutputFile,compress = TRUE)
return(df)
} else {
stop(paste("ERROR: could not find any variables"),paste(VctTSconditions,collapse=","))
}
}
#' @export
CheckIfColumnsAreAvailable<-function(df,VctColumnames){
#TODO check if columns exists before running functions!!
return(0)
}
#' @export
Query_TSAmin_masterset<-function(dfmaster_TSDEATHMEDICD10_visitdtonly,StrTrait,StrDescription,StrName,VctTSAcolumns,visitdt,StataOutputFile){
#dfmaster_TSDEATHMEDICD10_visitdtonly<-UKbioDatasetm
#StrTrait<-"a1def"
#StrDescription<-"asdsad"
#VctTSAcolumns<-"n_22152_"
#visitdt<-paste("ts_53_",2,"_0",sep="")
#StataOutputFile<-"/data_work/databases/ukbiobanks/dataset/"
### CHECK IF FILE EXISTS:
if( file.exists(StataOutputFile )) {
print(paste(StataOutputFile," ... already exists!, skipping"))
return(0)
}
print(paste(StrTrait,":",StrDescription))
dfmaster_TSDEATHMEDICD10_visitdtonly_subset<-dfmaster_TSDEATHMEDICD10_visitdtonly[ , grepl( paste(c("n_eid",visitdt,VctTSAcolumns), collapse='|') , names( dfmaster_TSDEATHMEDICD10_visitdtonly ) ) ]
### BirthDate:
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$Birthdate<- as.Date(paste('01', paste(dfmaster_TSDEATHMEDICD10_visitdtonly$n_52_0_0,dfmaster_TSDEATHMEDICD10_visitdtonly$n_34_0_0)), format='%d %m %Y')
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$Age<-(dfmaster_TSDEATHMEDICD10_visitdtonly_subset[,visitdt] - dfmaster_TSDEATHMEDICD10_visitdtonly_subset$Birthdate) /365.25
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAOveralMin<-NA ## set initial value
if( ncol(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)==5) {stop("Missing TSA column? cannot find anything; so at least 1 missing")} ## 0 column ; error; missing column.
for (VctTSAcolumn in VctTSAcolumns){
print(paste("mean-minimum age across columns:",VctTSAcolumn))
TSAcolumns_TF=grepl( paste(c(VctTSAcolumn), collapse='|') , names( dfmaster_TSDEATHMEDICD10_visitdtonly_subset ) )
if(length(which(TSAcolumns_TF))==1){ ## if only one visit is available no mean is possible; then duplicate it.
print("Only one visit is available.")
dfmaster_TSDEATHMEDICD10_visitdtonly_subset[paste(VctTSAcolumn,"tmp_0",sep="")]<-dfmaster_TSDEATHMEDICD10_visitdtonly_subset[TSAcolumns_TF]
TSAcolumns_TF=grepl( paste(c(VctTSAcolumn), collapse='|') , names( dfmaster_TSDEATHMEDICD10_visitdtonly_subset ) )
}
dfmaster_TSDEATHMEDICD10_visitdtonly_subset[TSAcolumns_TF][dfmaster_TSDEATHMEDICD10_visitdtonly_subset[TSAcolumns_TF] == -1]<-NA
dfmaster_TSDEATHMEDICD10_visitdtonly_subset[TSAcolumns_TF][dfmaster_TSDEATHMEDICD10_visitdtonly_subset[TSAcolumns_TF] == -3]<-NA
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAmin<-rowMins(as.matrix( dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ ,grepl(paste(VctTSAcolumn, collapse='|'),names(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)) ]),na.rm=TRUE)
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAmax<-rowMaxs(as.matrix( dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ ,grepl(paste(VctTSAcolumn, collapse='|'),names(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)) ]),na.rm=TRUE)
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSADiff<-dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAmax-dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAmin
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAmean<-rowMeans(as.matrix( dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ ,grepl(paste(VctTSAcolumn, collapse='|'),names(dfmaster_TSDEATHMEDICD10_visitdtonly_subset))]),na.rm=TRUE)
### filter outliers: difference between visits >5years.
outliers<-dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSADiff>5 & !is.infinite(dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSADiff)
if(length(which(outliers))>0){## check if outliers exist.
print("removing TSAd outliers based on age-difference>5")
dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ outliers,]$TSAmean<-NA
}
#dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ is.infinite(dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAmean),]$TSAmean<-NA
### add minimum to existing column;
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAOveralMin<-rowMins(as.matrix( cbind(dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAmean,dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAOveralMin)),na.rm=TRUE)
dfmaster_TSDEATHMEDICD10_visitdtonly_subset[is.infinite(dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAOveralMin),]$TSAOveralMin<-NA
}
dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd<-as.numeric(round((dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAOveralMin - dfmaster_TSDEATHMEDICD10_visitdtonly_subset$Age)*365.25))
# dfmaster_TSDEATHMEDICD10_visitdtonly_subset<-dfmaster_TSDEATHMEDICD10_visitdtonly_subset[dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAmean>=0,] # filter age>0?
dfmaster_TSDEATHMEDICD10_visitdtonly_subset<-(dfmaster_TSDEATHMEDICD10_visitdtonly_subset[!is.na(dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd),c("n_eid","TSAd")])
### Recalculate to Days from Reference.
dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ abs(dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd)<365.25 & dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd>0, "TSAd"] <- dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ abs(dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd)<365.25 & dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd>0, "TSAd"] /2 # within 1 year of diagnosis
dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ abs(dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd)<365.25 & dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd<0, "TSAd"] <- dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ abs(dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd)<365.25 & dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd<0, "TSAd"] /2 # within 1 year of diagnosis
dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd <= -365.25 ,"TSAd"]<-dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd <= -365.25 ,"TSAd"]+round(365.25/2) ## beyond one year in history
dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd >= 365.25 ,"TSAd"]<-dfmaster_TSDEATHMEDICD10_visitdtonly_subset[ dfmaster_TSDEATHMEDICD10_visitdtonly_subset$TSAd >= 365.25 ,"TSAd"]-round(365.25/2) ## beyond one year in future
if( is.data.frame(dfmaster_TSDEATHMEDICD10_visitdtonly_subset) && nrow(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)>0){
names(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)<-c("n_eid",paste(StrTrait,StrName,sep="_"))
attr(dfmaster_TSDEATHMEDICD10_visitdtonly_subset, "var.labels") <- c("Identifier",StrDescription)
attr(dfmaster_TSDEATHMEDICD10_visitdtonly_subset, "var.labels")<-substr(attr(dfmaster_TSDEATHMEDICD10_visitdtonly_subset, "var.labels"),0,80)
save.dta13(dfmaster_TSDEATHMEDICD10_visitdtonly_subset, StataOutputFile,compress = TRUE)
return(dfmaster_TSDEATHMEDICD10_visitdtonly_subset)
} else {
print(paste("WARNING: df empty; could not find any variables OR observations (could be due to visit 2 having no observations vs visit 1), maybe check: ",StrTrait,"///",paste(VctTSAcolumn,collapse=",")))
}
}
#' #' Extracting Vars From larger UK Bio dataset
#' #'
#' #' Extracts vars from ukbiobank from larger set of variables based on selection.
#' #'
#' #' @param UKbioDataset df
#' #' @param dfDefinitions_VarSelection df
#' #' @param StataOutputFile file
#' #' @keywords ExtractVarsFromMasterSet CreateUKBiobankPhentoypes
#' #' @return None
#' #'
#' #' @examples
#' #' ExtractVarsFromMasterSet(UKbioDataset,dfDefinitions_VarSelection,StataOutputFile)
#' #'
#' #' @export
#' ExtractVarsFromMasterSet<-function(UKbioDataset,dfDefinitions_VarSelection,StataOutputFile){
#'
#' #OutputdirMerged=paste(Outputdir,"/merged/",sep="")
#' #StataOutputFile=paste(OutputdirMerged,"/DataSelection.dta",sep="")
#'
#' ############################################################
#' ############ EXTRACT SELECTION FROM MASTERSET:
#' dfSQL_TraitTable_VarSelection_Selected<-dfDefinitions_VarSelection[dfDefinitions_VarSelection$Include==1,]
#' #as.vector(dfSQL_TraitTable_VarSelection_Selected$Var)
#' ### FOUND:
#' KeyAvail=names(UKbioDataset) %in% as.vector(dfSQL_TraitTable_VarSelection_Selected$var)
#' Avail=(names(UKbioDataset)[KeyAvail])
#' dfmasterSelection<-UKbioDataset[KeyAvail]
#' attr(dfmasterSelection,"var.labels")<-attr(UKbioDataset,"var.labels")[KeyAvail]
#' attr(dfmasterSelection, "var.labels")<-substr(attr(dfmasterSelection, "var.labels"),0,80)
#'
#' ### NOT FOUND
#' NotAvail=as.vector(dfSQL_TraitTable_VarSelection_Selected$Var[!as.vector(dfSQL_TraitTable_VarSelection_Selected$var) %in% names(UKbioDataset) ])
#'
#' print(paste("FOUND:",paste(Avail,collapse=", ")))
#' print(paste("MISSING:",paste(NotAvail,collapse=", ")))
#'
#' save.dta13(dfmasterSelection,StataOutputFile,compress = TRUE)
#'
#' }
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.