Name-based wrapper around read_recent_snapshots: resolves the
snapshot database from the dataset's configuration exactly as
run_dq_check does (the merged global + dataset config's
snapshot_db, defaulting to "data/snapshots.sqlite"), so the
whole workflow can be driven by dataset names without ever hand-typing a
database path. The returned id column is the snapshot id that
compare_snapshots takes to compare a specific pair of runs.
Value
The data frame documented in read_recent_snapshots:
one row per run, most recent first, empty (with the full column schema)
when the database does not exist or holds no rows for the dataset.
Note
As with run_dq_check, a relative snapshot_db
resolves against the R process's working directory, not against
config_dir. Call from the deployment root or use an absolute path
in the config, or the lookup will (like a run started from the wrong
directory) point at a database that isn't the deployment's.
Examples
tmp <- gsub("\\\\", "/", tempdir())
writeLines(paste0('snapshot_db: "', tmp, '/snap.sqlite"'),
file.path(tmp, "dqcheckr.yml"))
writeLines(c('dataset_name: "demo"', 'format: csv'),
file.path(tmp, "demo.yml"))
list_runs("demo", config_dir = tmp) # empty frame: no runs recorded yet
#> [1] id dataset_name
#> [3] run_timestamp file_name
#> [5] row_count col_count
#> [7] check_pass_count check_warn_count
#> [9] check_fail_count check_info_count
#> [11] overall_status new_cols_vs_previous
#> [13] missing_cols_vs_previous new_cols_vs_schema
#> [15] missing_cols_vs_schema comparison_mode
#> [17] render_status type_changed_cols_vs_previous
#> [19] report_file
#> <0 rows> (or 0-length row.names)