Verified Dataset Import

Roger He

2026-07-20

Before importing datasets into a report, it is useful to verify whether the data you have locally already matches what is on the platform. This avoids unnecessary re-imports and helps detect when data has changed. The functions described in this article provide a checksum-based workflow to compare local datasets against those already imported into a report, then the platform will only import what has changed.

Dataset format

All functions in this workflow accept datasets in the same list-of-lists format:

datasets <- list(
  list(
    data           = my_dataframe,        # R dataframe
    data_set_key   = "my_table_data",     # unique key used on the platform
    chart_types    = list("table"),       # determines CSV vs JSON serialization
    subgroups      = list("sex", "age"),  # optional: subgroup column names
    data_set_label = "My Table Data"      # optional: display label on the platform
  ),
  list(
    data           = my_nested_list,      # nested list for JSON-based chart types
    data_set_key   = "my_sankey_data",
    chart_types    = list("sankey"),
    subgroups      = list("priori_mi", "age_cat"),
    data_set_label = "My Sankey Data"
  )
)

chart_types controls how the data is serialized when computing checksums and during import: "sankey" and "sunburst" are written as JSON; all other chart types are written as CSV.

Step 1: Compare checksums

compare_dataset_checksums() computes SHA-256 checksums for your local datasets, fetches the checksums already stored on the platform, and compares them key by key.

match_result <- nswpr::compare_dataset_checksums(
  report_id    = "p0000",
  datasets     = datasets,
  version_type = "draft"   # optional: "draft", "published", or omitted for latest
)

The result is a named list — one entry per dataset key — each containing:

• local — SHA-256 checksum of your local data
• platform — SHA-256 checksum stored on the platform (NA if the dataset is not yet on the platform)
• match — TRUE if the checksums are identical

Choosing a version_type

The version_type parameter controls which version of the report’s datasets is used for comparison on the platform:

• "draft" — compares against the latest draft version. Use this when you are actively updating data and want to verify what is currently staged for the next publish.
• "published" — compares against the latest published version. Use this to check what end users are currently seeing.
• NA (default) — compares against the overall latest version regardless of publish status, equivalent to "draft" in most cases.

# Compare against the published version
match_result <- nswpr::compare_dataset_checksums("p0000", datasets, version_type = "published")

# Compare against the draft version
match_result <- nswpr::compare_dataset_checksums("p0000", datasets, version_type = "draft")

Step 2: Inspect the comparison

checksum_comparison_table() formats the result as a data frame for easy reading:

nswpr::checksum_comparison_table(match_result)

Example output:

dataset_key	local_checksum	platform_checksum	match
my_table_data	a3f1...	a3f1...	TRUE
my_sankey_data	b9c2...	(not on platform)	FALSE

A match of FALSE means the local dataset doesn’t match the one on the platform. A platform checksum of (not on platform) means it has never been imported before (identified by the dataset_key).

Step 3: Import

import_datasets() submits all data specified in datasets to the platform, which skips unnecessary imports automatically. Pass the match_result from the previous step — this avoids recomputing checksums and is recommended. If match_result is omitted it runs the comparison internally.

nswpr::import_datasets(
  report_id    = "p0000",
  datasets     = datasets,
  match_result = match_result
)

Full example

datasets <- list(
  list(
    data           = my_dataframe,
    data_set_key   = "my_table_data",
    chart_types    = list("table"),
    subgroups      = list("sex", "age"),
    data_set_label = "My Table Data"
  ),
  list(
    data           = my_nested_list,
    data_set_key   = "my_sankey_data",
    chart_types    = list("sankey"),
    subgroups      = list("priori_mi", "age_cat"),
    data_set_label = "My Sankey Data"
  )
)

match_result <- nswpr::compare_dataset_checksums("p0000", datasets)

nswpr::checksum_comparison_table(match_result)

nswpr::import_datasets("p0000", datasets, match_result)

Platform UI after import

Dataset details

Once a dataset has been imported, the platform displays the details per dataset including the checksum that was computed at import time.

Import history

The import history shows the status of every dataset specified in datasets, including those that were already up to date and not re-imported.