diff --git a/NAMESPACE b/NAMESPACE
index 69fe2334..356898c1 100644
--- a/NAMESPACE
+++ b/NAMESPACE
@@ -10,6 +10,7 @@ export(get_closest_matches)
 export(get_non_matched_elements)
 export(map_ds_spatial_units)
 export(map_spatial)
+export(read_write_metadata_tables)
 export(spatial_mapping_country)
 export(stattab_find_match)
 export(stattab_make_pattern)
diff --git a/R/metadata_tables.R b/R/metadata_tables.R
new file mode 100644
index 00000000..195ff3b8
--- /dev/null
+++ b/R/metadata_tables.R
@@ -0,0 +1,80 @@
+#' Read/Write metadata tables as csv
+#'
+#' @param ds the dataset: it is expected to have
+#'           ds$postgres_export
+#'           the language is to be expected the language
+#'           the ds is in (one language per pipeline)
+#' @param overwrite default is FALSE, when set to TRUE
+#'                  the metadata file will be overwritten
+#'
+#' @return metadata a tibble containing the metadata that
+#'                  matches what is on file
+#'                  the function watches for files:
+#'                  metadata_tables.csv
+#'                  metadata_table_columns.csv
+#'                  If they are not there or overwrite is set
+#'                  to TRUE templates for these files will be
+#'                  set up. Otherwise the files will be read
+#'                  and reported back
+#' @export
+#'
+#' @examples
+#' \dontrun{
+#'   # command to call when the metadata should be read or
+#'   # created: it will be created if it does not exist yet
+#'   read_write_metadata_tables(ds)
+#'
+#'   # command to call when the metadata should be overwritten
+#'   read_write_metadata_tables(ds, overwrite = TRUE)
+#' }
+read_write_metadata_tables <- function(ds, overwrite = FALSE) {
+
+  # set path for metadata storage
+  path_table <- paste0(ds$dir, "metadata_tables.csv")
+  path_table_columns <- paste0(ds$dir, "metadata_table_columns.csv")
+
+  if (file.exists(path_table) && file.exists(path_table_columns) &&!overwrite) {
+    metadata_tables = readr::read_csv(path_table)
+    metadata_table_columns = readr::read_csv(path_table_columns)
+    return(list(metadata_tables=metadata_tables,
+                metadata_table_columns=metadata_table_columns))
+  }
+
+  # make template for metadata files
+
+  # metadata_tables entry: if possible load metadata
+  table_description <- ""
+  if (ds$organization == "bfs" && ds$format == "px") {
+    bfs_metadata <- BFS::bfs_get_metadata(number_bfs = ds$sheet, language = ds$lang)
+    table_description <- unique(bfs_metadata$title)
+  }
+  metadata_tables <- tibble::tibble(
+    name = ds$name,
+    language = ds$lang,
+    description = table_description
+  )
+
+  # columns are received from the columns of the postgres export
+  columns <- colnames(ds$postgres_export)
+  columns <- columns[!(columns %in% c("jahr", "spatialunit_uid"))]
+  columns_count <- length(columns)
+  metadata_table_columns <- tibble::tibble(
+    name = columns,
+    table_name = rep(ds$name, times = columns_count),
+    data_type = rep("", times = columns_count),
+    description = rep("", times = columns_count)
+  )
+
+  # write the tibble to a file
+  write.csv(metadata_tables, path_table, row.names = FALSE, quote = FALSE)
+  write.csv(metadata_table_columns, path_table_columns, row.names = FALSE, quote = FALSE)
+  print(
+    paste(
+      "metadata templates written to\n",
+      ds$dir,
+      "\nPlease complete metadata for pipeline."
+    )
+  )
+  return(list(metadata_tables=metadata_tables,
+              metadata_table_columns=metadata_table_columns))
+}
diff --git a/man/read_write_metadata_tables.Rd b/man/read_write_metadata_tables.Rd
new file mode 100644
index 00000000..027c33e6
--- /dev/null
+++ b/man/read_write_metadata_tables.Rd
@@ -0,0 +1,41 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/metadata_tables.R
+\name{read_write_metadata_tables}
+\alias{read_write_metadata_tables}
+\title{Read/Write metadata tables as csv}
+\usage{
+read_write_metadata_tables(ds, overwrite = FALSE)
+}
+\arguments{
+\item{ds}{the dataset: it is expected to have
+ds$postgres_export
+the language is to be expected the language
+the ds is in (one language per pipeline)}
+
+\item{overwrite}{default is FALSE, when set to TRUE
+the metadata file will be overwritten}
+}
+\value{
+metadata a tibble containing the metadata that
+matches what is on file
+the function watches for files:
+metadata_tables.csv
+metadata_table_columns.csv
+If they are not there or overwrite is set
+to TRUE templates for these files will be
+set up. Otherwise the files will be read
+and reported back
+}
+\description{
+Read/Write metadata tables as csv
+}
+\examples{
+\dontrun{
+  # command to call when the metadata should be read or
+  # created: it will be created if it does not exist yet
+  read_write_metadata_tables(ds)
+
+  # command to call when the metadata should be overwritten
+  read_write_metadata_tables(ds, overwrite = TRUE)
+}
+}
diff --git a/pipelines/A6/A6.R b/pipelines/A6/A6.R
index 116ba75a..b02b0e4d 100644
--- a/pipelines/A6/A6.R
+++ b/pipelines/A6/A6.R
@@ -36,8 +36,8 @@ ds$postgres_export
 
 # -------------------------------------------------------------------------
 # Step: Testrun queries on sqllite
-#   input: ds$postgres_export, A6/queries.sql
-#   output: A6/queries.log
+#   input: ds$postgres_export, pipelines/A6/queries.sql
+#   output: pipelines/A6/queries.log
 # -------------------------------------------------------------------------
 
 statbotData::testrun_queries(
@@ -45,3 +45,11 @@ statbotData::testrun_queries(
   ds$dir,
   ds$name
 )
+
+# -------------------------------------------------------------------------
+# Step: Write metadata tables
+#   input: ds$postgres_export
+#   output: pipelines/A6/metadata.csv
+# -------------------------------------------------------------------------
+
+read_write_metadata_tables(ds)
diff --git a/pipelines/A6/metadata_table_columns.csv b/pipelines/A6/metadata_table_columns.csv
new file mode 100644
index 00000000..dabb903b
--- /dev/null
+++ b/pipelines/A6/metadata_table_columns.csv
@@ -0,0 +1,3 @@
+name,table_name,data_type,description
+thema,abstimmungsvorlagen_seit_1971,,
+anzahl_abstimmungsvorlagen,abstimmungsvorlagen_seit_1971,,
diff --git a/pipelines/A6/metadata_tables.csv b/pipelines/A6/metadata_tables.csv
new file mode 100644
index 00000000..b7f96bc9
--- /dev/null
+++ b/pipelines/A6/metadata_tables.csv
@@ -0,0 +1,2 @@
+name,language,description
+abstimmungsvorlagen_seit_1971,de,Volksabstimmungen (Anzahl Vorlagen nach Thema seit 1971)