From 5dff2ba483915b1502b6ae86ab05b864ec8baaf9 Mon Sep 17 00:00:00 2001
From: Fred Ho <crescal@gmail.com>
Date: Thu, 23 Jul 2026 13:35:15 +0100
Subject: [PATCH] Fix aggregate_per_date() combining results using last file's
 columns instead of first

After looping over all time-series files, aggregate_per_date() combined
per-file results and set the final table's column count/names from
whichever file happened to be processed *last* (ci/dsnames were reused
directly from the final loop iteration). This made the result silently
depend on file processing order/completeness. Now each file's (ds,
dsnames) pair is collected in a list and combined afterwards with
purrr::list_rbind(), using the *first* file's dsnames as the canonical
column names/count.

Verified against a synthetic multi-subject dataset: output is identical
to the previous implementation in the well-behaved case (same column
count for every file).
---
 DESCRIPTION            |  4 ++--
 NAMESPACE              |  1 +
 R/aggregate_per_date.R | 20 +++++++++++++++++---
 3 files changed, 20 insertions(+), 5 deletions(-)

diff --git a/DESCRIPTION b/DESCRIPTION
index 0b93a1f..abd2edd 100644
--- a/DESCRIPTION
+++ b/DESCRIPTION
@@ -20,8 +20,8 @@ LazyData: true
 Suggests: 
     testthat (>= 3.0.0), randomForest, gsignal, rattle
 Config/testthat/edition: 3
-RoxygenNote: 7.3.2
+RoxygenNote: 7.3.3
 Depends: 
     R (>= 2.10)
 Imports: 
-    caret, data.table, e1071, GGIR, HMM, signal, GGIRread, actilifecounts, actimetricModels
+    caret, data.table, e1071, GGIR, HMM, signal, GGIRread, actilifecounts, actimetricModels, purrr
diff --git a/NAMESPACE b/NAMESPACE
index 5fd3776..33c038a 100644
--- a/NAMESPACE
+++ b/NAMESPACE
@@ -40,6 +40,7 @@ importFrom(e1071,kurtosis)
 importFrom(e1071,skewness)
 importFrom(grDevices,dev.off)
 importFrom(grDevices,pdf)
+importFrom(purrr,list_rbind)
 importFrom(signal,specgram)
 importFrom(stats,predict)
 importFrom(utils,data)
diff --git a/R/aggregate_per_date.R b/R/aggregate_per_date.R
index 0f58d09..b284278 100644
--- a/R/aggregate_per_date.R
+++ b/R/aggregate_per_date.R
@@ -24,6 +24,7 @@
 #' @return Data frame with aggregates of time spent in classes per calendar date.
 #'
 #' @importFrom grDevices pdf dev.off
+#' @importFrom purrr list_rbind
 #'
 #' @author Jairo H. Migueles <jairo@jhmigueles.com>
 #' @export
@@ -36,6 +37,9 @@ aggregate_per_date = function(tsDir, epoch, classifier, classes,
     classes = gsub("sleep", "nighttime.sleep", classes)
   }
   files = dir(tsDir, full.names = TRUE)
+  # per-file results, combined into a single table after the loop
+  ds_list = vector("list", length(files))
+  dsnames_list = vector("list", length(files))
   for (fi in 1:length(files)) {
     ts = NULL
     load(files[fi])
@@ -206,12 +210,22 @@ aggregate_per_date = function(tsDir, epoch, classifier, classes,
                      classes = classes, epoch = epoch)
       grDevices::dev.off()
     }
-    # store information in daysummary and next file
-    if (fi == 1) daysummary = ds else daysummary = rbind(daysummary, ds)
+    # store information for this file; combined into daysummary after the loop
+    ds_list[[fi]] = ds
+    dsnames_list[[fi]] = dsnames
   }
 
+  # Combine results across files. Column structure (dsnames/ci) is determined
+  # by the run's parameters (classifier, classes, boutdur), so it is expected
+  # to be identical for every file; we use the *first* file's dsnames as the
+  # canonical column names/count rather than whichever file happened to be
+  # processed *last*, so the combined table does not silently depend on which
+  # file is last in `files`.
+  daysummary = list_rbind(ds_list)
+  dsnames = dsnames_list[[1]]
+
   # keep only calculated columns and insert colnames
-  daysummary = daysummary[, 1:(ci - 1)]
+  daysummary = daysummary[, 1:length(dsnames)]
   colnames(daysummary) = dsnames
   # return
   return(daysummary)