diff --git a/CRAN-SUBMISSION b/CRAN-SUBMISSION
index dbc50b70..e096d9c0 100644
--- a/CRAN-SUBMISSION
+++ b/CRAN-SUBMISSION
@@ -1,3 +1,3 @@
-Version: 0.1.3
-Date: 2024-01-19 22:06:35 UTC
-SHA: a13cec24750e0efdda2624199e8c4e48506e287b
+Version: 0.1.4
+Date: 2024-05-03 19:26:08 UTC
+SHA: ef7c70ebd2dc070c696af0fb4d24c066bbcef487
diff --git a/DESCRIPTION b/DESCRIPTION
index bb8fc8f8..f251268b 100644
--- a/DESCRIPTION
+++ b/DESCRIPTION
@@ -1,6 +1,6 @@
 Package: aorsf
 Title: Accelerated Oblique Random Forests
-Version: 0.1.4
+Version: 0.1.4.9001
 Authors@R: c(
     person("Byron", "Jaeger", , "bjaeger@wakehealth.edu", role = c("aut", "cre"),
            comment = c(ORCID = "0000-0001-7399-2299")),
diff --git a/NEWS.md b/NEWS.md
index 5b38cc32..0e16427e 100644
--- a/NEWS.md
+++ b/NEWS.md
@@ -1,3 +1,7 @@
+# aorsf 0.1.5 (unreleased)
+
+* fixed an issue where omitting NA values would cause an error in regression forests.
+
 # aorsf 0.1.4
 
 * `orsf_vs` now returns a column that contains non-reference coded variable names (see https://github.com/ropensci/aorsf/pull/52).
diff --git a/R/orsf_R6.R b/R/orsf_R6.R
index 29979b46..e7352e68 100644
--- a/R/orsf_R6.R
+++ b/R/orsf_R6.R
@@ -441,10 +441,6 @@ ObliqueForest <- R6::R6Class(
    # object and then use this function. We need checks for that case.
    new_data <- new_data %||% self$data
 
-   if(oobag && nrow(new_data) != self$n_obs){
-    stop("input data must have ", self$n_obs, " observations to ",
-         "compute out-of-bag predictions.", call. = FALSE)
-   }
 
    # run checks before you assign new values to object.
    # otherwise, if a check throws an error, the object will
@@ -476,6 +472,12 @@ ObliqueForest <- R6::R6Class(
 
      private$init_data_rows_complete()
      private$prep_x()
+
+     if(oobag && nrow(private$x) != self$n_obs){
+      stop("input data must have ", self$n_obs, " observations to ",
+           "compute out-of-bag predictions.", call. = FALSE)
+     }
+
      # y and w do not need to be prepped for prediction,
      # but they need to match orsf_cpp()'s expectations
      private$y <- matrix(0, nrow = 1, ncol = 1)
@@ -3153,7 +3155,7 @@ ObliqueForest <- R6::R6Class(
    private$y <- select_cols(self$data, private$data_names$y)
 
    if(self$na_action == 'omit' && !placeholder)
-    private$y <- private$y[private$data_rows_complete, ]
+    private$y <- private$y[private$data_rows_complete, , drop = FALSE]
 
    private$prep_y_internal(placeholder)
 
@@ -3877,7 +3879,6 @@ ObliqueForestSurvival <- R6::R6Class(
 
   clean_pred_oobag_internal = function(){
 
-
    # put the oob predictions into the same order as the training data.
    unsorted <- collapse::radixorder(private$data_row_sort)
    self$pred_oobag <- self$pred_oobag[unsorted, , drop = FALSE]
@@ -3939,7 +3940,10 @@ ObliqueForestSurvival <- R6::R6Class(
 
    # must sort if oobag b/c when oobag_rows were originally created,
    # it was after the data had been sorted.
-   if(oobag) private$sort_inputs(sort_y = FALSE)
+   if(oobag){
+    private$sort_inputs(sort_y = FALSE)
+    unsorted <- collapse::radixorder(private$data_row_sort)
+   }
 
    cpp_args = private$prep_cpp_args(x = private$x,
                                     y = private$y,
@@ -3963,6 +3967,12 @@ ObliqueForestSurvival <- R6::R6Class(
 
      results[[i]] <- do.call(orsf_cpp, args = cpp_args)$pred_new
 
+     if(oobag){
+      # put the oob predictions into the same order as the training data.
+      results[[i]] <- results[[i]][unsorted, , drop = FALSE]
+     }
+
+
      results[[i]] <- private$clean_pred_new(results[[i]])
 
     }
@@ -3983,9 +3993,7 @@ ObliqueForestSurvival <- R6::R6Class(
 
    if(oobag){
     # put the oob predictions into the same order as the training data.
-    unsorted <- collapse::radixorder(private$data_row_sort)
     out_values <- out_values[unsorted, , drop = FALSE]
-
    }
 
    private$clean_pred_new(out_values)
diff --git a/cran-comments.md b/cran-comments.md
index 50c2d321..66bf139d 100644
--- a/cran-comments.md
+++ b/cran-comments.md
@@ -1,6 +1,2 @@
 
-With the help of a colleague, I was able to reproduce the undefined behavior error. I am confident that it is now resolved.
-
-I understand you are taking an extended break from reviewing this package and I'm comfortable waiting until the break is concluded. Thank you for everything you do.
-
-I saw that it took a long time to run tests on my last submission. I added rules to skip tests on CRAN that require longer run times. I hope this is helpful.
+This patch release contains a few minor bug fixes.
diff --git a/tests/testthat/test-na_action.R b/tests/testthat/test-na_action.R
new file mode 100644
index 00000000..3049fb2e
--- /dev/null
+++ b/tests/testthat/test-na_action.R
@@ -0,0 +1,113 @@
+
+test_that(
+ desc = "na action of omit works with oobag preds",
+ code = {
+
+  mtcars_na <- mtcars
+
+  mtcars_na$vs <- factor(mtcars_na$vs)
+
+  mtcars_na$disp[1] <- NA
+  mtcars_na$hp[3] <- NA
+
+  regr_fit <- orsf(
+   data = mtcars_na, formula = mpg ~ .,
+   n_tree = n_tree_test,
+   tree_seeds = seeds_standard,
+   na_action = 'omit'
+  )
+
+  clsf_fit <- orsf(
+   data = mtcars_na, formula = vs ~ .,
+   n_tree = n_tree_test,
+   tree_seeds = seeds_standard,
+   na_action = 'omit'
+  )
+
+  surv_fit <- orsf(
+   mtcars_na, mpg + vs ~ .,
+   n_tree = n_tree_test,
+   tree_seeds = seeds_standard,
+   na_action = 'omit'
+  )
+
+  expect_equal(nrow(na.omit(mtcars_na)), regr_fit$n_obs)
+  expect_equal(nrow(na.omit(mtcars_na)), clsf_fit$n_obs)
+  expect_equal(nrow(na.omit(mtcars_na)), surv_fit$n_obs)
+
+  clsf_prd_oob <- predict(clsf_fit, oobag = TRUE)
+  regr_prd_oob <- predict(regr_fit, oobag = TRUE)
+  surv_prd_oob <- predict(surv_fit, oobag = TRUE)
+
+  na_rows <- which(is.na(clsf_fit$pred_oobag))
+  expect_true(all(is.na(clsf_prd_oob[na_rows, drop = FALSE])))
+
+  na_rows <- which(is.na(regr_fit$pred_oobag))
+  expect_true(all(is.na(regr_prd_oob[na_rows, drop = FALSE])))
+
+  na_rows <- which(is.na(surv_fit$pred_oobag))
+  expect_true(all(is.na(surv_prd_oob[na_rows, drop = FALSE])))
+
+ })
+
+
+
+test_that(
+ desc = "na action of impute works with oobag preds",
+ code = {
+
+  mtcars_na <- mtcars
+
+  mtcars_na$vs <- factor(mtcars_na$vs)
+
+  mtcars_na$disp[1] <- NA
+  mtcars_na$hp[3] <- NA
+
+  regr_fit <- orsf(
+   data = mtcars_na, formula = mpg ~ .,
+   n_tree = n_tree_test,
+   tree_seeds = seeds_standard,
+   na_action = 'impute_meanmode'
+  )
+
+  clsf_fit <- orsf(
+   data = mtcars_na, formula = vs ~ .,
+   n_tree = n_tree_test,
+   tree_seeds = seeds_standard,
+   na_action = 'impute_meanmode'
+  )
+
+  surv_fit <- orsf(
+   mtcars_na, mpg + vs ~ .,
+   n_tree = n_tree_test,
+   tree_seeds = seeds_standard,
+   na_action = 'impute_meanmode'
+  )
+
+  expect_equal(nrow(mtcars_na), regr_fit$n_obs)
+  expect_equal(nrow(mtcars_na), clsf_fit$n_obs)
+  expect_equal(nrow(mtcars_na), surv_fit$n_obs)
+
+  clsf_prd_oob <- predict(clsf_fit, oobag = TRUE)
+  regr_prd_oob <- predict(regr_fit, oobag = TRUE)
+  surv_prd_oob <- predict(surv_fit, oobag = TRUE)
+
+  expect_equal(nrow(mtcars_na), regr_fit$n_obs)
+  expect_equal(nrow(mtcars_na), clsf_fit$n_obs)
+  expect_equal(nrow(mtcars_na), surv_fit$n_obs)
+
+  clsf_prd_oob <- predict(clsf_fit, oobag = TRUE)
+  regr_prd_oob <- predict(regr_fit, oobag = TRUE)
+  surv_prd_oob <- predict(surv_fit, oobag = TRUE)
+
+  na_rows <- which(is.na(clsf_fit$pred_oobag))
+  expect_true(all(is.na(clsf_prd_oob[na_rows, drop = FALSE])))
+
+  na_rows <- which(is.na(regr_fit$pred_oobag))
+  expect_true(all(is.na(regr_prd_oob[na_rows, drop = FALSE])))
+
+  na_rows <- which(is.na(surv_fit$pred_oobag))
+  expect_true(all(is.na(surv_prd_oob[na_rows, drop = FALSE])))
+
+ })
+