Drugs and formularies

FDA FAERS — Adverse-Event Drug Involvement (DRUG file)

drug_silver.fact_faers_drug·live·silver·FDA FAERS — Adverse-Event Drug Involvement (DRUG file)

Rows

11,131,453

Columns

History

N/A

Card

nocolumns only

Typed tools

What it is

1FDA FAERS (Adverse Event Reporting System) DRUG file — per adverse-event report × drug: the drug's role (PS primary-suspect / SS secondary / C concomitant / I interacting), drug name, active ingredient, NDA number, route/dose, dechallenge/ rechallenge.
2The drug-safety surveillance signal — links adverse events to the drug substrate (join drugname/prod_ai/nda_num to dim_drug / dim_drug_application). ~1.6-2.2M rows/quarter;
32025Q1-2026Q2 loaded (six quarters).

Grain

Key	Grain
N/A	1One row per line of a quarter's FAERS DRUG file: physical key (report_quarter, row_seq), row_seq = the line's position in that quarter's file (accepted 2026-09-23). 2(report_quarter, primaryid, drug_seq) is NOT a key: the publisher repeats it (21,797 duplicate groups / 68,550 rows in 2025Q1). 3Quarters are noncumulative and case-versioned; count cases with the current-case rule in the context doc, never rows or distinct primaryid.

Key

Grain

N/A

1One row per line of a quarter's FAERS DRUG file: physical key (report_quarter, row_seq), row_seq = the line's position in that quarter's file (accepted 2026-09-23).
2(report_quarter, primaryid, drug_seq) is NOT a key: the publisher repeats it (21,797 duplicate groups / 68,550 rows in 2025Q1).
3Quarters are noncumulative and case-versioned; count cases with the current-case rule in the context doc, never rows or distinct primaryid.

Columns (35)

Column	Type	Meaning
`report_quarter`	string
`row_seq`	bigint
`primaryid`	string
`caseid`	string
`role_cod`	string
`drugname`	string
`prod_ai`	string
`val_vbm`	string
`route`	string
`dose_vbm`	string
`cum_dose_chr`	string
`cum_dose_unit`	string
`dechal`	string
`rechal`	string
`nda_num`	string
`dose_amt`	string
`dose_unit`	string
`dose_form`	string
`dose_freq`	string
`drug_seq`	double
`observed_period`	date
`asserted_at`	date
`effective_period_start`	date
`effective_period_end`	date
`cms_vintage_sequence`	int
`parse_revision`	int
`parser_artifact_hash`	string
`source_url`	string
`sha256_content_hash`	string
`source_id`	string
`is_current`	boolean
`bronze_object_key`	string
`s3_version_id`	string
`vintage_tag`	string
`drug_source_version_id`	string

Query it

describe_data_source(source_fqn="drug_silver.fact_faers_drug")     # the card, live
query_substrate(sql="SELECT * FROM drug_silver.fact_faers_drug WHERE <filter> LIMIT 100")

Item	Value
Typed tools that read it	none; raw SQL
Identity key	`report_quarter,row_seq`
Join	`how_to_join(source_fqn="drug_silver.fact_faers_drug")`
Raw door	Athena (read-only via query_substrate): drug_silver.fact_faers_drug
Note	Iceberg drug_silver.fact_faers_drug, partition by report_quarter; bulk-load (append_buffered 300k -> flush_buffered_replace_partitions); a rerun of a loaded quarter skips on the ZIP SHA-256 recorded in medicare_source_registry_cache.

This page as markdown: /docs-md/data/drug_silver_fact_faers_drug.md · every page in one file: /llms-full.txt