diff --git a/NAMESPACE b/NAMESPACE
index e71084b8d..a0a1db0be 100644
--- a/NAMESPACE
+++ b/NAMESPACE
@@ -70,7 +70,6 @@ importFrom(data.table,.SD)
 importFrom(data.table,CJ)
 importFrom(data.table,as.IDate)
 importFrom(data.table,as.data.table)
-importFrom(data.table,copy)
 importFrom(data.table,data.table)
 importFrom(data.table,dcast)
 importFrom(data.table,fcase)
diff --git a/R/check.R b/R/check.R
index 29d419e66..313a169e2 100644
--- a/R/check.R
+++ b/R/check.R
@@ -31,10 +31,9 @@ check_quantiles <- function(posterior, req_probs = c(0.5, 0.95, 0.2, 0.8)) {
 #' @return a copy `data.table` version of `obs` with `report_date` and
 #' `reference_date` as [IDateTime] format.
 #'
-#' @importFrom data.table as.data.table
 #' @family check
 check_dates <- function(obs) {
-  obs <- data.table::as.data.table(obs)
+  obs <- coerce_dt(obs)
   if (is.null(obs$reference_date) && is.null(obs$report_date)) {
     stop(
       "Both reference_date and report_date must be present in order to use this
@@ -109,9 +108,9 @@ check_by <- function(obs, by = NULL) {
 
 #' Add a reserved grouping variable if missing
 #'
-#' @param x A data.table
+#' @param x A `data.table`, optionally with a `.group` variable
 #'
-#' @return A data table with a `.group` variable
+#' @return `x`, definitely with a `.group` variable
 #' @family check
 add_group <- function(x) {
   if (is.null(x[[".group"]])) {
@@ -167,3 +166,37 @@ check_modules_compatible <- function(modules) {
   }
   return(invisible(NULL))
 }
+
+#' @title Coerce `data.table`s
+#'
+#' @description Provides consistent coercion of inputs to [data.table]
+#' with error handling
+#'
+#' @param data any of the types supported by [data.table::as.data.table()]
+#'
+#' @param new logical; if `TRUE` (default), a new `data.table` is returned
+#'
+#' @param required_cols character vector of required columns
+#' 
+#' @param forbidden_cols character vector of forbidden columns
+#' 
+#' @return a `data.table`; if `data` is a `data.table`, the returned object
+#' will have a new address, unless `new = FALSE`.
+#' i.e. be distinct from the original and not cause any side effects with
+#' changes.
+#'
+#' @details TODO
+#'
+#' @importFrom data.table as.data.table
+#' @family utils
+coerce_dt <- function(
+  data, required_cols, forbidden_cols, new = TRUE
+) {
+  if (!new && inherits(data, "data.table")) {
+    dt <- data
+  } else {
+    dt <- data.table::as.data.table(data)
+  }
+
+  return(dt[])
+}
\ No newline at end of file
diff --git a/R/formula-tools.R b/R/formula-tools.R
index 791ec2bee..8fcf3e9cf 100644
--- a/R/formula-tools.R
+++ b/R/formula-tools.R
@@ -30,7 +30,6 @@
 #'
 #' @inheritParams enw_formula
 #' @family formulatools
-#' @importFrom data.table copy
 #' @importFrom stats as.formula
 #' @export
 #' @examples
@@ -39,7 +38,7 @@
 enw_manual_formula <- function(data, fixed = NULL, random = NULL,
                                custom_random = NULL, no_contrasts = FALSE,
                                add_intercept = TRUE) {
-  data <- data.table::copy(data)
+  data <- coerce_dt(data)
   if (add_intercept) {
     form <- "1"
   } else {
@@ -286,7 +285,6 @@ rw <- function(time, by, type = c("independent", "dependent")) {
 #' formula.
 #'  - `effects`: A `data.frame` describing the random effect structure of the
 #' new effects.
-#' @importFrom data.table copy
 #' @family formulatools
 #' @examples
 #' data <- enw_example("preproc")$metareference[[1]]
@@ -298,7 +296,7 @@ construct_rw <- function(rw, data) {
   if (!inherits(rw, "enw_rw_term")) {
     stop("rw must be a random walk term as constructed by rw")
   }
-  data <- data.table::copy(data)
+  data <- coerce_dt(data)
 
   if (!is.numeric(data[[rw$time]])) {
     stop(
@@ -318,7 +316,7 @@ construct_rw <- function(rw, data) {
   )
   ctime <- paste0("c", rw$time)
   terms <- grep(ctime, colnames(data), value = TRUE)
-  fdata <- data.table::copy(data)
+  fdata <- coerce_dt(data)
   fdata <- fdata[, c(terms, rw$by), with = FALSE]
   if (!is.null(rw$by)) {
     if (is.null(fdata[[rw$by]])) {
@@ -407,7 +405,6 @@ re <- function(formula) {
 #' converted into one.
 #'
 #' @family formulatools
-#' @importFrom data.table as.data.table
 #' @importFrom purrr map
 #' @examples
 #' # Simple examples
@@ -429,7 +426,7 @@ construct_re <- function(re, data) {
   if (!inherits(re, "enw_re_term")) {
     stop("re must be a random effect term as constructed by re")
   }
-  data <- data.table::as.data.table(data)
+  data <- coerce_dt(data)
 
   # extract random and fixed effects
   fixed <- strsplit(re$fixed, " + ", fixed = TRUE)[[1]]
@@ -597,7 +594,7 @@ construct_re <- function(re, data) {
 #' @family formulatools
 #' @export
 #' @importFrom purrr map transpose
-#' @importFrom data.table as.data.table rbindlist setnafill
+#' @importFrom data.table rbindlist setnafill
 #' @examples
 #' # Use meta data for references dates from the Germany COVID-19
 #' # hospitalisation data.
@@ -625,7 +622,7 @@ construct_re <- function(re, data) {
 #' # to specify an independent random effect per strata.
 #' enw_formula(~ (1 + day | week:month), data = data)
 enw_formula <- function(formula, data, sparse = TRUE) {
-  data <- data.table::as.data.table(data)
+  data <- coerce_dt(data)
 
   # Parse formula
   parsed_formula <- parse_formula(formula)
diff --git a/R/model-design-tools.R b/R/model-design-tools.R
index 2aeb47713..5d2eb4ff7 100644
--- a/R/model-design-tools.R
+++ b/R/model-design-tools.R
@@ -58,7 +58,6 @@ mod_matrix <- function(formula, data, sparse = TRUE, ...) {
 #' @return A list containing the formula, the design matrix, and the index.
 #' @family modeldesign
 #' @export
-#' @importFrom data.table as.data.table
 #' @importFrom stats terms contrasts model.matrix
 #' @importFrom purrr map
 #' @examples
@@ -71,7 +70,7 @@ mod_matrix <- function(formula, data, sparse = TRUE, ...) {
 enw_design <- function(formula, data, no_contrasts = FALSE, sparse = TRUE,
                        ...) {
   # make data.table and copy
-  data <- data.table::as.data.table(data)
+  data <- coerce_dt(data)
 
   # make all character variables factors
   chars <- colnames(data)[sapply(data, is.character)]
@@ -227,7 +226,7 @@ enw_add_pooling_effect <- function(effects, var_name = "sd",
 #' metaobs <- data.frame(week = 1:3, .group = c(1,1,2))
 #' enw_add_cumulative_membership(metaobs, "week")
 enw_add_cumulative_membership <- function(metaobs, feature) {
-  metaobs <- data.table::as.data.table(metaobs)
+  metaobs <- coerce_dt(metaobs)
   metaobs <- add_group(metaobs)
 
   cfeature <- paste0("c", feature)
diff --git a/R/model-module-helpers.R b/R/model-module-helpers.R
index 65cf5b952..cd3030d71 100644
--- a/R/model-module-helpers.R
+++ b/R/model-module-helpers.R
@@ -10,7 +10,7 @@
 #' @family modelmodulehelpers
 enw_reps_with_complete_refs <- function(new_confirm, max_delay, by = NULL) {
   check_by(new_confirm, by = by)
-  rep_with_complete_ref <- data.table::as.data.table(new_confirm)
+  rep_with_complete_ref <- coerce_dt(new_confirm)
   rep_with_complete_ref <- rep_with_complete_ref[,
     .(n = .N),
     by = c(by, "report_date")
@@ -38,7 +38,7 @@ enw_reps_with_complete_refs <- function(new_confirm, max_delay, by = NULL) {
 enw_reference_by_report <- function(missing_reference, reps_with_complete_refs,
                                     metareference, max_delay) {
   # Make a complete data.frame of all possible reference and report dates
-  miss_lk <- data.table::copy(metareference)[
+  miss_lk <- coerce_dt(metareference)[
     ,
     .(reference_date = date, .group)
   ]
diff --git a/R/model-modules.R b/R/model-modules.R
index a1cd19f46..f4c55a503 100644
--- a/R/model-modules.R
+++ b/R/model-modules.R
@@ -425,7 +425,7 @@ enw_expectation <- function(r = ~ 0 + (1 | day:.group), generation_time = 1,
 #' @inheritParams enw_obs
 #' @inheritParams enw_formula
 #' @family modelmodules
-#' @importFrom data.table setorderv copy dcast
+#' @importFrom data.table setorderv dcast
 #' @importFrom purrr map
 #' @export
 #' @examples
@@ -469,16 +469,16 @@ enw_missing <- function(formula = ~1, data) {
     )
 
     # Get the indexes for when grouped observations start and end
-    miss_lookup <- data.table::copy(rep_w_complete_ref)
+    miss_lookup <- coerce_dt(rep_w_complete_ref)
     data_list$miss_st <- miss_lookup[, n := seq_len(.N), by = ".group"]
     data_list$miss_st <- data_list$miss_st[, .(n = max(n)), by = ".group"]$n
     data_list$miss_cst <- miss_lookup[, n := seq_len(.N)]
     data_list$miss_cst <- data_list$miss_cst[, .(n = max(n)), by = ".group"]$n
 
     # Get (and order) reported cases with a missing reference date
-    missing_reference <- data.table::copy(data$missing_reference[[1]])
+    missing_reference <- coerce_dt(data$missing_reference[[1]])
     data.table::setkeyv(missing_reference, c(".group", "report_date"))
-    data_list$missing_reference <- data.table::copy(missing_reference)[
+    data_list$missing_reference <- coerce_dt(missing_reference)[
       rep_w_complete_ref,
       on = c("report_date", ".group")
     ][, confirm]
@@ -567,7 +567,7 @@ enw_obs <- function(family = c("negbin", "poisson"), data) {
   latest_matrix <- latest_obs_as_matrix(data$latest[[1]])
 
   # get new confirm for processing
-  new_confirm <- data.table::copy(data$new_confirm[[1]])
+  new_confirm <- coerce_dt(data$new_confirm[[1]])
   data.table::setkeyv(new_confirm, c(".group", "reference_date", "delay"))
 
   # get flat observations
diff --git a/R/model-tools.R b/R/model-tools.R
index 74ea16303..8a4787c5e 100644
--- a/R/model-tools.R
+++ b/R/model-tools.R
@@ -82,14 +82,13 @@ enw_formula_as_data_list <- function(formula, prefix,
 #' two vector (specifying the mean and standard deviation of the prior).
 #' @family modeltools
 #' @inheritParams enw_replace_priors
-#' @importFrom data.table copy
 #' @importFrom purrr map
 #' @export
 #' @examples
 #' priors <- data.frame(variable = "x", mean = 1, sd = 2)
 #' enw_priors_as_data_list(priors)
 enw_priors_as_data_list <- function(priors) {
-  priors <- data.table::as.data.table(priors)
+  priors <- coerce_dt(priors)
   priors[, variable := paste0(variable, "_p")]
   priors <- priors[, .(variable, mean, sd)]
   priors <- split(priors, by = "variable", keep.by = FALSE)
@@ -120,7 +119,6 @@ enw_priors_as_data_list <- function(priors) {
 #' @return A data.table of prior definitions (variable, mean and sd).
 #' @family modeltools
 #' @export
-#' @importFrom data.table as.data.table
 #' @examples
 #' # Update priors from a data.frame
 #' priors <- data.frame(variable = c("x", "y"), mean = c(1, 2), sd = c(1, 2))
@@ -142,16 +140,13 @@ enw_priors_as_data_list <- function(priors) {
 #'
 #' enw_replace_priors(default_priors, fit_priors)
 enw_replace_priors <- function(priors, custom_priors) {
-  custom_priors <- data.table::as.data.table(custom_priors)[
+  custom_priors <- coerce_dt(custom_priors)[
     ,
-    .(variable, mean = as.numeric(mean), sd = as.numeric(sd))
-  ]
-  custom_priors <- custom_priors[
-    ,
-    variable := gsub("\\[([^]]*)\\]", "", variable)
+    .(variable = gsub("\\[([^]]*)\\]", "", variable),
+      mean = as.numeric(mean), sd = as.numeric(sd))
   ]
   variables <- custom_priors$variable
-  priors <- data.table::as.data.table(priors)[!(variable %in% variables)]
+  priors <- coerce_dt(priors)[!(variable %in% variables)]
   priors <- rbind(priors, custom_priors, fill = TRUE)
   return(priors[])
 }
diff --git a/R/model-validation.R b/R/model-validation.R
index 4dce6f529..71189e16f 100644
--- a/R/model-validation.R
+++ b/R/model-validation.R
@@ -28,7 +28,7 @@
 #'
 #' @return A `data.table` as returned by [scoringutils::score()].
 #' @family modelvalidation
-#' @importFrom data.table copy setnames
+#' @importFrom data.table setnames
 #' @export
 #' @examplesIf interactive()
 #' library(data.table)
@@ -60,7 +60,7 @@ enw_score_nowcast <- function(nowcast, latest_obs, log = FALSE,
   if (!is.null(long_nowcast[["mad"]])) {
     long_nowcast[, "mad" := NULL]
   }
-  latest_obs <- data.table::copy(latest_obs)
+  latest_obs <- coerce_dt(latest_obs)
   data.table::setnames(latest_obs, "confirm", "true_value", skip_absent = TRUE)
   latest_obs[, report_date := NULL]
   cols <- intersect(colnames(nowcast), colnames(latest_obs))
diff --git a/R/plot.R b/R/plot.R
index 9d0a354bb..b8cb817d1 100644
--- a/R/plot.R
+++ b/R/plot.R
@@ -32,7 +32,6 @@ enw_plot_theme <- function(plot) {
 #'
 #' @family plot
 #' @importFrom scales comma
-#' @importFrom data.table copy
 #' @export
 #' @examples
 #' nowcast <- enw_example("nowcast")
@@ -53,7 +52,7 @@ enw_plot_obs <- function(obs, latest_obs = NULL, log = TRUE, ...) {
     )
 
   if (!is.null(latest_obs)) {
-    latest_obs <- data.table::copy(latest_obs)
+    latest_obs <- coerce_dt(latest_obs)
     latest_obs[, latest_confirm := confirm]
     plot <- plot +
       geom_point(
@@ -151,7 +150,7 @@ enw_plot_nowcast_quantiles <- function(nowcast, latest_obs = NULL,
 #' enw_plot_pp_quantiles(nowcast) +
 #'  ggplot2::facet_wrap(ggplot2::vars(reference_date), scales = "free")
 enw_plot_pp_quantiles <- function(pp, log = FALSE, ...) {
-  pp <- data.table::copy(pp)
+  pp <- coerce_dt(pp)
   pp[, confirm := new_confirm]
   plot <- enw_plot_quantiles(
     pp,
diff --git a/R/postprocess.R b/R/postprocess.R
index f2a3b29be..3fbbf063d 100644
--- a/R/postprocess.R
+++ b/R/postprocess.R
@@ -78,7 +78,7 @@ enw_posterior <- function(fit, variables = NULL,
 #' @inheritParams enw_posterior
 #' @family postprocess
 #' @export
-#' @importFrom data.table as.data.table copy setorderv
+#' @importFrom data.table setorderv
 #' @examples
 #' fit <- enw_example("nowcast")
 #' enw_nowcast_summary(fit$fit[[1]], fit$latest[[1]])
@@ -94,7 +94,7 @@ enw_nowcast_summary <- function(fit, obs,
 
   max_delay <- nrow(nowcast) / max(obs$.group)
 
-  ord_obs <- data.table::copy(obs)
+  ord_obs <- coerce_dt(obs)
   ord_obs <- ord_obs[reference_date > (max(reference_date) - max_delay)]
   data.table::setorderv(ord_obs, c(".group", "reference_date"))
   nowcast <- cbind(
@@ -121,7 +121,7 @@ enw_nowcast_summary <- function(fit, obs,
 #' @inheritParams enw_nowcast_summary
 #' @family postprocess
 #' @export
-#' @importFrom data.table as.data.table copy setorderv
+#' @importFrom data.table setorderv
 #' @examples
 #' fit <- enw_example("nowcast")
 #' enw_nowcast_samples(fit$fit[[1]], fit$latest[[1]])
@@ -138,7 +138,7 @@ enw_nowcast_samples <- function(fit, obs) {
   )
   max_delay <- nrow(nowcast) / (max(obs$.group) * max(nowcast$.draw))
 
-  ord_obs <- data.table::copy(obs)
+  ord_obs <- coerce_dt(obs)
   ord_obs <- ord_obs[reference_date > (max(reference_date) - max_delay)]
   data.table::setorderv(ord_obs, c(".group", "reference_date"))
   ord_obs <- data.table::data.table(
@@ -230,14 +230,14 @@ enw_summarise_samples <- function(samples, probs = c(
 #' added.
 #' @family postprocess
 #' @export
-#' @importFrom data.table as.data.table setcolorder
+#' @importFrom data.table setcolorder
 #' @examples
 #' fit <- enw_example("nowcast")
 #' obs <- enw_example("obs")
 #' nowcast <- summary(fit, type = "nowcast")
 #' enw_add_latest_obs_to_nowcast(nowcast, obs)
 enw_add_latest_obs_to_nowcast <- function(nowcast, obs) {
-  obs <- data.table::as.data.table(obs)
+  obs <- coerce_dt(obs)
   obs <- add_group(obs)
   obs <- obs[, .(reference_date, .group, latest_confirm = confirm)]
   out <- merge(
@@ -266,7 +266,7 @@ enw_add_latest_obs_to_nowcast <- function(nowcast, obs) {
 #' @inheritParams enw_posterior
 #' @family postprocess
 #' @export
-#' @importFrom data.table as.data.table copy setorderv
+#' @importFrom data.table setorderv
 #' @examples
 #' fit <- enw_example("nowcast")
 #' enw_pp_summary(fit$fit[[1]], fit$new_confirm[[1]], probs = c(0.5))
@@ -280,7 +280,7 @@ enw_pp_summary <- function(fit, diff_obs,
     probs = probs
   )
 
-  ord_obs <- data.table::copy(diff_obs)
+  ord_obs <- coerce_dt(diff_obs)
   data.table::setorderv(ord_obs, c(".group", "reference_date"))
   pp <- cbind(
     ord_obs,
diff --git a/R/preprocess.R b/R/preprocess.R
index f81465dd3..ec314fac5 100644
--- a/R/preprocess.R
+++ b/R/preprocess.R
@@ -23,7 +23,6 @@
 #' @family preprocess
 #' @importFrom data.table setkeyv
 #' @export
-#' @importFrom data.table as.data.table
 #' @examples
 #' obs <- data.frame(
 #'  reference_date = as.Date("2021-01-01"),
@@ -33,14 +32,14 @@
 enw_metadata <- function(obs, target_date = c(
                            "reference_date", "report_date"
                          )) {
-  obs <- data.table::as.data.table(obs)
+  obs <- coerce_dt(obs)
   choices <- eval(formals()$target_date)
   target_date <- match.arg(target_date)
   date_to_drop <- setdiff(choices, target_date)
 
   obs <- add_group(obs)
 
-  metaobs <- setnames(data.table::as.data.table(obs), target_date, "date")
+  metaobs <- setnames(coerce_dt(obs), target_date, "date")
   suppressWarnings(
     metaobs[
       ,
@@ -121,7 +120,7 @@ enw_add_metaobs_features <- function(metaobs,
                                      holidays_to = "Sunday",
                                      datecol = "date") {
   # localize and check metaobs input
-  metaobs <- data.table::as.data.table(metaobs)
+  metaobs <- coerce_dt(metaobs)
   if (is.null(metaobs[[datecol]])) {
     stop(sprintf("metaobs does not have datecol '%s'.", datecol))
   }
@@ -214,7 +213,7 @@ enw_add_metaobs_features <- function(metaobs,
 #'
 #' @family preprocess
 #' @export
-#' @importFrom data.table copy data.table rbindlist setkeyv
+#' @importFrom data.table data.table rbindlist setkeyv
 #' @importFrom purrr map
 #' @examples
 #' metaobs <- data.frame(date = as.Date("2021-01-01") + 0:4)
@@ -230,7 +229,7 @@ enw_extend_date <- function(metaobs, days = 20, direction = "end") {
   } else {
     filt_fn <- max
   }
-  metaobs <- data.table::as.data.table(metaobs)
+  metaobs <- coerce_dt(metaobs)
   metaobs <- add_group(metaobs)
   exts <- metaobs[, .SD[date == filt_fn(date)], by = .group]
   exts <- split(exts, by = ".group")
@@ -269,7 +268,6 @@ enw_extend_date <- function(metaobs, days = 20, direction = "end") {
 #'
 #' @family preprocess
 #' @export
-#' @importFrom data.table as.data.table copy
 #' @examples
 #' obs <- data.frame(x = 1:3, y = 1:3)
 #' enw_assign_group(obs)
@@ -281,10 +279,10 @@ enw_assign_group <- function(obs, by = NULL) {
       "from your data before calling `enw_assign_group`."
     )
   }
-  check_by(obs, by = by)
-  obs <- data.table::as.data.table(obs)
+  check_by(obs, by = by) # TODO: order? should ensure a dt first or ...?
+  obs <- coerce_dt(obs)
   if (length(by) != 0) {
-    groups_index <- data.table::copy(obs)
+    groups_index <- coerce_dt(obs)
     groups_index <- unique(groups_index[, ..by])
     groups_index[, .group := seq_len(.N)]
     obs <- merge(obs, groups_index, by = by, all.x = TRUE)
@@ -305,7 +303,6 @@ enw_assign_group <- function(obs, by = NULL) {
 #' @inheritParams enw_cumulative_to_incidence
 #' @family preprocess
 #' @export
-#' @importFrom data.table as.data.table copy
 #' @examples
 #' obs <- data.frame(report_date = as.Date("2021-01-01") + -2:0)
 #' obs$reference_date <- as.Date("2021-01-01")
@@ -332,7 +329,6 @@ enw_add_delay <- function(obs) {
 #' @inheritParams enw_latest_data
 #' @family preprocess
 #' @export
-#' @importFrom data.table copy
 #' @examples
 #' obs <- data.frame(report_date = as.Date("2021-01-01") + 0:2)
 #' obs$reference_date <- as.Date("2021-01-01")
@@ -580,12 +576,11 @@ enw_incidence_to_cumulative <- function(obs, by = NULL) {
 #' @inheritParams enw_preprocess_data
 #' @family preprocess
 #' @export
-#' @importFrom data.table copy
 #' @examples
 #' obs <- enw_example("preprocessed")$obs[[1]]
 #' enw_delay_filter(obs, max_delay = 2)
 enw_delay_filter <- function(obs, max_delay) {
-  obs <- data.table::as.data.table(obs)
+  obs <- coerce_dt(obs)
   obs <- add_group(obs)
   obs <- obs[,
     .SD[
@@ -612,12 +607,12 @@ enw_delay_filter <- function(obs, max_delay) {
 #' being observations by reporting delay.
 #' @family preprocess
 #' @export
-#' @importFrom data.table as.data.table dcast setorderv
+#' @importFrom data.table dcast setorderv
 #' @examples
 #' obs <- enw_example("preprocessed")$new_confirm
 #' enw_reporting_triangle(obs)
 enw_reporting_triangle <- function(obs) {
-  obs <- data.table::as.data.table(obs)
+  obs <- coerce_dt(obs)
   obs <- add_group(obs)
   if (any(obs$new_confirm < 0)) {
     warning(
@@ -648,7 +643,7 @@ enw_reporting_triangle <- function(obs) {
 #' rt <- enw_reporting_triangle(obs)
 #' enw_reporting_triangle_to_long(rt)
 enw_reporting_triangle_to_long <- function(obs) {
-  obs <- data.table::as.data.table(obs)
+  obs <- coerce_dt(obs)
   obs <- add_group(obs)
   reports_long <- data.table::melt(
     obs,
@@ -675,7 +670,7 @@ enw_reporting_triangle_to_long <- function(obs) {
 #'
 #' @inheritParams enw_preprocess_data
 #' @export
-#' @importFrom data.table as.data.table CJ
+#' @importFrom data.table CJ
 #' @family preprocess
 #' @examples
 #' obs <- data.frame(
@@ -685,7 +680,7 @@ enw_reporting_triangle_to_long <- function(obs) {
 #' enw_complete_dates(obs)
 enw_complete_dates <- function(obs, by = NULL, max_delay,
                                missing_reference = TRUE) {
-  obs <- data.table::as.data.table(obs)
+  obs <- coerce_dt(obs)
   obs <- check_dates(obs)
   check_group(obs)
   check_by(obs)
@@ -751,7 +746,6 @@ enw_complete_dates <- function(obs, by = NULL, max_delay,
 #' group.
 #'
 #' @export
-#' @importFrom data.table as.data.table
 #' @family preprocess
 #' @examples
 #' obs <- data.frame(
@@ -966,7 +960,7 @@ enw_construct_data <- function(obs, new_confirm, latest, missing_reference,
 #' @family preprocess
 #' @inheritParams enw_cumulative_to_incidence
 #' @export
-#' @importFrom data.table as.data.table data.table
+#' @importFrom data.table data.table
 #' @examples
 #' library(data.table)
 #'
diff --git a/man/add_group.Rd b/man/add_group.Rd
index 484eca339..38216b6a0 100644
--- a/man/add_group.Rd
+++ b/man/add_group.Rd
@@ -7,10 +7,10 @@
 add_group(x)
 }
 \arguments{
-\item{x}{A data.table}
+\item{x}{A \code{data.table}, optionally with a \code{.group} variable}
 }
 \value{
-A data table with a \code{.group} variable
+\code{x}, definitely with a \code{.group} variable
 }
 \description{
 Add a reserved grouping variable if missing
diff --git a/man/check_dates.Rd b/man/check_dates.Rd
index 6f1ae9143..6376152b7 100644
--- a/man/check_dates.Rd
+++ b/man/check_dates.Rd
@@ -11,8 +11,8 @@ check_dates(obs)
 \code{reference_date} columns.}
 }
 \value{
-Returns the input \code{data.frame} with dates converted to date format
-if not already.
+a copy \code{data.table} version of \code{obs} with \code{report_date} and
+\code{reference_date} as \link{IDateTime} format.
 }
 \description{
 Check Report and Reference Dates are present
diff --git a/man/coerce_date.Rd b/man/coerce_date.Rd
index 727093c48..2c956eaa0 100644
--- a/man/coerce_date.Rd
+++ b/man/coerce_date.Rd
@@ -38,6 +38,7 @@ tryCatch(
 }
 \seealso{
 Utility functions
+\code{\link{coerce_dt}()},
 \code{\link{convert_cmdstan_to_rstan}()},
 \code{\link{expose_stan_fns}()},
 \code{\link{is.Date}()},
diff --git a/man/coerce_dt.Rd b/man/coerce_dt.Rd
new file mode 100644
index 000000000..d15ecc201
--- /dev/null
+++ b/man/coerce_dt.Rd
@@ -0,0 +1,39 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/check.R
+\name{coerce_dt}
+\alias{coerce_dt}
+\title{Coerce \code{data.table}s}
+\usage{
+coerce_dt(data, required_cols, forbidden_cols, new = TRUE)
+}
+\arguments{
+\item{data}{any of the types supported by \code{\link[data.table:as.data.table]{data.table::as.data.table()}}}
+
+\item{required_cols}{character vector of required columns}
+
+\item{forbidden_cols}{character vector of forbidden columns}
+
+\item{new}{logical; if \code{TRUE} (default), a new \code{data.table} is returned}
+}
+\value{
+a \code{data.table}; if \code{data} is a \code{data.table}, the returned object
+will have a new address, unless \code{new = FALSE}.
+i.e. be distinct from the original and not cause any side effects with
+changes.
+}
+\description{
+Provides consistent coercion of inputs to \link{data.table}
+with error handling
+}
+\details{
+TODO
+}
+\seealso{
+Utility functions
+\code{\link{coerce_date}()},
+\code{\link{convert_cmdstan_to_rstan}()},
+\code{\link{expose_stan_fns}()},
+\code{\link{is.Date}()},
+\code{\link{stan_fns_as_string}()}
+}
+\concept{utils}
diff --git a/man/convert_cmdstan_to_rstan.Rd b/man/convert_cmdstan_to_rstan.Rd
index 311a8f283..4fecc97d1 100644
--- a/man/convert_cmdstan_to_rstan.Rd
+++ b/man/convert_cmdstan_to_rstan.Rd
@@ -19,6 +19,7 @@ Convert Cmdstan to Rstan
 \seealso{
 Utility functions
 \code{\link{coerce_date}()},
+\code{\link{coerce_dt}()},
 \code{\link{expose_stan_fns}()},
 \code{\link{is.Date}()},
 \code{\link{stan_fns_as_string}()}
diff --git a/man/expose_stan_fns.Rd b/man/expose_stan_fns.Rd
index d9354e20d..36174e2e4 100644
--- a/man/expose_stan_fns.Rd
+++ b/man/expose_stan_fns.Rd
@@ -28,6 +28,7 @@ make use of this function apart from when exploring package functionality.
 \seealso{
 Utility functions
 \code{\link{coerce_date}()},
+\code{\link{coerce_dt}()},
 \code{\link{convert_cmdstan_to_rstan}()},
 \code{\link{is.Date}()},
 \code{\link{stan_fns_as_string}()}
diff --git a/man/is.Date.Rd b/man/is.Date.Rd
index a1532129c..9e0f57bf9 100644
--- a/man/is.Date.Rd
+++ b/man/is.Date.Rd
@@ -18,6 +18,7 @@ Checks that an object is a date
 \seealso{
 Utility functions
 \code{\link{coerce_date}()},
+\code{\link{coerce_dt}()},
 \code{\link{convert_cmdstan_to_rstan}()},
 \code{\link{expose_stan_fns}()},
 \code{\link{stan_fns_as_string}()}
diff --git a/man/stan_fns_as_string.Rd b/man/stan_fns_as_string.Rd
index 37895301d..9703b02ed 100644
--- a/man/stan_fns_as_string.Rd
+++ b/man/stan_fns_as_string.Rd
@@ -21,6 +21,7 @@ Read in a stan function file as a character string
 \seealso{
 Utility functions
 \code{\link{coerce_date}()},
+\code{\link{coerce_dt}()},
 \code{\link{convert_cmdstan_to_rstan}()},
 \code{\link{expose_stan_fns}()},
 \code{\link{is.Date}()}
diff --git a/tests/testthat/helper-functions.R b/tests/testthat/helper-functions.R
index 172ecba6d..9b4565281 100644
--- a/tests/testthat/helper-functions.R
+++ b/tests/testthat/helper-functions.R
@@ -29,4 +29,12 @@ round_numerics <- function(dt) {
   cols <- colnames(dt)[purrr::map_lgl(dt, is.numeric)]
   dt <- dt[, (cols) := lapply(.SD, round, 0), .SDcols = cols]
   return(dt)
-}
\ No newline at end of file
+}
+
+dt_copies <- function(...) {
+  lapply(list(...), data.table::copy)
+}
+
+dt_compare_all <- function(ref_copies, ...) {
+  all(mapply(function(l, r) all(l == r), ref_copies, list(...)))
+}
diff --git a/tests/testthat/test-datatable-disconnection.R b/tests/testthat/test-datatable-disconnection.R
new file mode 100644
index 000000000..a13219f12
--- /dev/null
+++ b/tests/testthat/test-datatable-disconnection.R
@@ -0,0 +1,21 @@
+
+test_that("`add_group` maintains the same `data.table` object", {
+  dummy <- data.table::data.table(dummy = 1:10)
+  dummy_addr <- data.table::address(dummy)
+  add_group(dummy)
+  expect_equal(dummy_addr, data.table::address(dummy))
+  dummy <- add_group(dummy)
+  expect_equal(dummy_addr, data.table::address(dummy))
+})
+
+test_that("`coerce_dt` gives new `data.table` object", {
+  dummy <- data.table::data.table(dummy = 1:10)
+  newdt <- coerce_dt(dummy)
+  expect_false(data.table::address(newdt) == data.table::address(dummy))
+})
+
+test_that("`coerce_dt` gives new `data.table` object, unless asked not to", {
+  dummy <- data.table::data.table(dummy = 1:10)
+  newdt <- coerce_dt(dummy, new = FALSE)
+  expect_true(data.table::address(newdt) == data.table::address(dummy))
+})
diff --git a/tests/testthat/test-enw_replace_priors.R b/tests/testthat/test-enw_replace_priors.R
index 58a7c8fb4..5d100cf88 100644
--- a/tests/testthat/test-enw_replace_priors.R
+++ b/tests/testthat/test-enw_replace_priors.R
@@ -39,3 +39,13 @@ test_that("enw_replace_priors can replace default priors with those from an
     updated_priors[variable %in% variables]$sd, as.numeric(fit_priors$sd)
   )
 })
+
+test_that("enw_replace_priors does not modify input `data.table`s", {
+   priors <- data.table::data.table(variable = c("x", "y"), mean = c(1, 2), sd = c(1, 2))
+   custom_priors <- data.table::data.table(variable = "x[1]", mean = 10, sd = 2)
+   refs <- dt_copies(priors, custom_priors)
+   newpriors <- enw_replace_priors(priors, custom_priors)
+   expect_true(data.table::address(newpriors) != data.table::address(priors))
+   expect_true(data.table::address(newpriors) != data.table::address(custom_priors))
+   expect_true(dt_compare_all(refs, priors, custom_priors))
+ })