diff --git a/R/Start.R b/R/Start.R
index c741de2226c5184f93eaef261aa4743ae258d53e..7eb43af280d50ad19c03aacf4670872c6663be1f 100644
--- a/R/Start.R
+++ b/R/Start.R
@@ -893,7 +893,7 @@ Start <- function(..., # dim = indices/selectors,
                                           dim_params, dim_reorder_params)
 
   # Check if file pattern contains '$var$' substring
-  if (!grepl("$var$", dim_params[[found_pattern_dim]], fixed = TRUE)) {
+  if (any(!grepl("$var$", dim_params[[found_pattern_dim]], fixed = TRUE))) {
     .warning(paste0("The special wildcard '$var$' is not present in the file ",
                     "path. This might cause Start() to fail if it cannot parse",
                     "the inner dimensions in all the files."))
@@ -2503,6 +2503,10 @@ Start <- function(..., # dim = indices/selectors,
             }
             # Find the largest length of each time step
             inner_dim_lengths <- do.call(pmax, inner_dim_lengths_cat)
+            ## NOTE: NA values can be present if the size of a depending 
+            ## dimension varies along its depended dim. Removing them allows
+            ## retrieval of the common indices. Could cause other issues?
+            inner_dim_lengths <- inner_dim_lengths[which(!is.na(inner_dim_lengths))]
           }
 
           fri <- first_round_indices <- NULL
diff --git a/tests/testthat/test-Start-multiple_depends.R b/tests/testthat/test-Start-multiple_depends.R
new file mode 100644
index 0000000000000000000000000000000000000000..473d0614522458decee0aaae21a4d3bc36f72f20
--- /dev/null
+++ b/tests/testthat/test-Start-multiple_depends.R
@@ -0,0 +1,65 @@
+suppressMessages({
+# This unit test tests the case where a depended dimension has multiple
+# depending dimensions and the 'all' selector is used for a depending dim.
+
+path <- "/esarchive/exp/CMIP6/$dcpp$/HadGEM3-GC31-MM/DCPP/MOHC/HadGEM3-GC31-MM/$dcpp$/r1i1p1f2/Omon/tos/gn/v20200417/$var$_Omon_HadGEM3-GC31-MM_$dcpp$_s$sdate$-r1i1p1f2_gn_$chunk$.nc"
+
+path <- paste0('/esarchive/scratch/aho/startR_unittest_files', path)
+
+sdates <- c('2018', '2019')
+
+test_that("1. ", {
+suppressWarnings(
+dat1 <- Start(dat = path,
+              var = 'tos',
+              chunk = 'all', 
+              time = indices(1:14),
+              time_across = 'chunk',
+              sdate = sdates,
+              dcpp = list('2018' = "dcppA-hindcast", '2019' = "dcppB-forecast"),
+              dcpp_depends = 'sdate',
+              chunk_depends = 'sdate',
+              merge_across_dims = TRUE,
+              largest_dims_length = TRUE,
+              i = indices(450:460),
+              j = indices(685:700),
+              return_vars = list(time = c('chunk', 'sdate')),
+              retrieve = TRUE)
+)
+
+suppressWarnings(
+dat2 <- Start(dat = path,
+              var = 'tos',
+              chunk = list('2018' = c('201811-201812', '201901-201912'),
+                           '2019' = c('201911-201912', '202001-202012')),
+              time = 'all',
+              time_across = 'chunk',
+              sdate = sdates,
+              dcpp = list('2018' = "dcppA-hindcast", '2019' = "dcppB-forecast"),
+              dcpp_depends = 'sdate',
+              chunk_depends = 'sdate',
+              merge_across_dims = TRUE,
+              largest_dims_length = TRUE,
+              i = indices(450:460),
+              j = indices(685:700),
+              return_vars = list(time = c('chunk', 'sdate')),
+              retrieve = TRUE)
+)
+
+expect_equal(
+  as.vector(dat1),
+  as.vector(dat2)
+)
+expect_equal(
+  mean(dat2, na.rm = T),
+  29.21144,
+  tolerance = 0.0001
+)
+expect_equal(
+  dat1[1, 1, 2, 2, 1, 1:3, 10],
+  c(28.84955, 28.84827, 28.84126),
+  tolerance = 0.0001
+  )
+})
+
+}) #suppressMessages