Fix issue with non-default dayborder in part 5

NEWS.md

@@ -4,6 +4,8 @@
 
 - Part 1: Fix bug where on machines with GMT timezone and R >= 4.3.0, for GENEActiv .bin files, the starting timestamps of M$metalong and M$metashort were truncated to midninght #1000
 
+- Part 5: Fix bug where part5 output showed duplicate rows when argument dayborder is set to nondefault value #1006
+
 # CHANGES IN GGIR VERSION 3.0-2
 
 - Part 2: Fix bug that caused part 2 to struggle with corrupt ActiGraph .gt3x files #972

R/g.analyse.R

@@ -67,7 +67,7 @@
   n_ws2_perday = (1440*60) / ws2
   n_ws3_perday = (1440*60) / ws3
   if (((nrow(metalong)/((1440*60)/ws2)*10) - (nrow(metashort)/((60/ws3)*1440)) * 10) > 1) {
-    cat("Matrices 'metalong' and 'metashort' are not compatible")
+    stop("Matrices 'metalong' and 'metashort' are not compatible")
   }
   #----------------------
   # Extract ID centrally
@@ -168,9 +168,6 @@
   # LW = length(which(as.numeric(qcheck) != 1)) / (60/ws3) #number of minutes wear time between first and last midnights
   nfulldays = (lastmidnighti - firstmidnighti) / ((3600/ws2)*24)
   ndays = length(midnights) + 1 #ceiling(nfulldays + 2) # ceiling to cope with days with 23 hours
-  if (ndays != round(ndays)) { #day saving time causing trouble?
-    cat("One day in this measurement is longer or shorter than 24 hours (probably related to day saving time)")
-  }
   #-------------------------------------
   # Detect orientation (at the moment only desighned for hip with accelerometer):
   # assess which angle per axis is most strongly 24 hour correlated:
@@ -279,8 +276,6 @@
       average24h = average24h / average24hc
       AveAccAve24hr[h] = mean(average24h) #average acceleration in an average 24 hour cycle
     }
-  } else {
-    cat("file skipped for general average caculation because not enough data")
   }
   rm(metalong); rm(metashort)
   dataqual_summary = data.frame(clipping_score = LC2  / ((LD/1440)*96),

R/g.analyse.perday.R

@@ -30,12 +30,10 @@
     if (firstmidnighti == 1) {  #if measurement starts at midnight
       ndays = ndays - 1
       startatmidnight =  1
-      cat("measurement starts at midnight or there is no midnight")
     }
     if (lastmidnight == time[length(time)] & nrow(metashort) < ((60/ws3) * 1440)) {	#if measurement ends at midnight
       ndays = ndays - 1
       endatmidnight = 1
-      cat("measurement ends at midnight or there is no midnight")
     }
   }
 
@@ -141,7 +139,7 @@
     if (length(params_247[["qwindow"]]) < 2) params_247[["qwindow"]] = c()
     if (length(params_247[["qwindow"]]) > 0) {
       if (length(qwindow_names) == 1) {
-        cat("Argument to qwindow is invalid, requires a vector of at least length 2")
+        warning("Argument to qwindow is invalid, requires a vector of at least length 2")
       }
       if (length(qwindow_names) == 2) {
         if (params_247[["qwindow"]][1] != 0 | params_247[["qwindow"]][2] != 24) {

R/g.part5.definedays.R

@@ -34,7 +34,7 @@ g.part5.definedays = function(nightsi, wi, indjump, nightsi_bu,
     if (nightsi[1] == 1) {
       wi = wi + 1
       # add extra nightsi to get the last day processed (as wi has been increased by 1)
-      nightsi = c(nightsi, nightsi + (24*(60/epochSize) * 60) - 1)
+      nightsi = c(nightsi, nightsi[length(nightsi)] + (24*(60/epochSize) * 60))
     }
     if (length(nightsi) >= wi) {
       if (wi == 1) {
@@ -89,6 +89,7 @@ g.part5.definedays = function(nightsi, wi, indjump, nightsi_bu,
     # in MM, also define segments of the day based on qwindow
     if (!is.na(qqq[1]) & !is.na(qqq[2])) {
       fullQqq = qqq[1]:qqq[2]
+      firstepoch = format(ts$time[qqq[1]],  "%H:%M:%S")
       lastepoch = format(ts$time[qqq[2]],  "%H:%M:%S")
       qnames = NULL
       if (is.data.frame(qwindow)) {
@@ -108,7 +109,7 @@ g.part5.definedays = function(nightsi, wi, indjump, nightsi_bu,
       }
       # build up segments
       segments = list(qqq)
-      segments_timing = paste("00:00:00", lastepoch, sep = "-")
+      segments_timing = paste(firstepoch, lastepoch, sep = "-")
       segments_names = "MM"
       si = 2
       do.segments = TRUE

R/g.report.part5.R

@@ -119,8 +119,6 @@ g.report.part5 = function(metadatadir = c(), f0 = c(), f1 = c(), loglocation = c
         out = base::merge(tmp, out, all = TRUE)
       }
       if (length(tail_expansion_log) != 0) {
-        # col2na = grep(pattern = "spt|sleeplog_used",
-        #               x = names(out), value = FALSE)
         col2na = grep(pattern = paste0("sleep_efficiency|N_atleast5minwakenight|daysleeper|",
                                        "daysleeper|sleeplog_used|_spt_sleep|_spt_wake"),
                       x = names(out), value = FALSE)
@@ -146,10 +144,8 @@ g.report.part5 = function(metadatadir = c(), f0 = c(), f1 = c(), loglocation = c
 
     # order data.frame
     outputfinal$window_number = as.numeric(gsub(" ", "", outputfinal$window_number))
-    # outputfinal_segments = outputfinal[which(!outputfinal$window %in% c("MM", "WW")),]
-    # outputfinal = outputfinal[which(outputfinal$window %in% c("MM", "WW")),]
-    outputfinal = outputfinal[order(outputfinal$filename, outputfinal$window, outputfinal$window_number),]
-
+    outputfinal = outputfinal[order(outputfinal$filename, outputfinal$window_number, outputfinal$window), ]
+
     # split results to different spreadsheets in order to minimize individual
     # filesize and to ease organising dataset
     uwi = as.character(unique(outputfinal$window))
@@ -475,7 +471,7 @@ g.report.part5 = function(metadatadir = c(), f0 = c(), f1 = c(), loglocation = c
                   }
                   OF3tmp = OF3[, columns2keep]
                   foo34 = function(df,aggPerIndividual,nameold,namenew,cval,window) {
-                    # function to help with calculating additinal variables
+                    # function to help with calculating additional variables
                     # related to counting how many days of measurement there are
                     # that meet a certain criteria
                     # cval is a vector with 0 and 1, indicating whether the criteria is met