SEC-Financial-Statements-And-Notes-Dataset / sql /sps /usp_sec_get_filing_metadata.sql
DenyTranDFW's picture
Add Dockerfile, docker-compose, setup script, SQL stored procedures, schema docs, and indexes
af136d8 verified
USE sec_fsnds;
GO
/*
============================================================================
dbo.usp_sec_get_filing_metadata
============================================================================
Given an accession number (adsh), return everything you'd want to know
about the filing *before* pulling the statements themselves:
Result set 1: the sub row (company + filing metadata)
Result set 2: row counts per statement in pre (non-dimensional)
Result set 3: row counts per statement in num (dimn>0 vs dimn=0)
Result set 4: text fact count from txt
Parameters
----------
@adsh accession number
Notes
-----
* Text fact count (result set 4) will be slow on sec_fsnds unless you
add IX_txt_sub (see sql/indexes/sec_fsnds_recommended_indexes.sql).
Without that index the txt read is a full scan of 48M rows / 100 GB.
============================================================================
*/
IF OBJECT_ID('dbo.usp_sec_get_filing_metadata', 'P') IS NOT NULL
DROP PROCEDURE dbo.usp_sec_get_filing_metadata;
GO
CREATE PROCEDURE dbo.usp_sec_get_filing_metadata
@adsh VARCHAR(20)
AS
BEGIN
SET NOCOUNT ON;
DECLARE @subID INT = (SELECT subID FROM dbo.sub WHERE adsh = @adsh);
IF @subID IS NULL
BEGIN
RAISERROR('No sub row for adsh=%s', 16, 1, @adsh);
RETURN;
END;
-- Result 1: full sub row
SELECT
s.subID, s.adsh, s.cik, s.name, s.sic, s.form
,CONVERT(date, s.period) AS period
,s.filed, s.accepted, s.fy, s.fp, s.fye
,s.countryba, s.stprba, s.cityba, s.zipba
,s.ein, s.wksi, s.prevrpt, s.detail, s.instance
,s.pubfloatusd, s.floatdate
FROM dbo.sub s
WHERE s.subID = @subID;
-- Result 2: presentation row counts per statement
SELECT
ls.stmt
,ls.label
,COUNT(*) AS pre_rows
,SUM(CASE WHEN p.inpth = 0 THEN 1 ELSE 0 END) AS pre_rows_main
,SUM(CASE WHEN p.inpth = 1 THEN 1 ELSE 0 END) AS pre_rows_parenthetical
FROM dbo.pre p
INNER JOIN dbo.lkp_stmt ls ON ls.stmtID = p.stmtID
WHERE p.subID = @subID
GROUP BY ls.stmtID, ls.stmt, ls.label
ORDER BY ls.stmtID;
-- Result 3: numeric fact counts (dimensional vs flat)
SELECT
COUNT(*) AS num_rows_total
,SUM(CASE WHEN n.dimn = 0 THEN 1 ELSE 0 END) AS num_rows_flat
,SUM(CASE WHEN n.dimn > 0 THEN 1 ELSE 0 END) AS num_rows_dimensional
,SUM(CASE WHEN n.coregID IS NOT NULL THEN 1 ELSE 0 END) AS num_rows_coreg
,COUNT(DISTINCT n.tagID) AS distinct_tags
,MIN(n.ddate) AS min_ddate
,MAX(n.ddate) AS max_ddate
FROM dbo.num n
WHERE n.subID = @subID
OPTION (RECOMPILE);
-- Result 4: text facts
SELECT
COUNT(*) AS txt_rows_total
,COUNT(DISTINCT x.tagID) AS distinct_text_tags
,SUM(CAST(x.txtlen AS BIGINT)) AS total_text_length_bytes
FROM dbo.txt x
WHERE x.subID = @subID;
END;
GO
/*
============================================================================
Example call
============================================================================
EXEC dbo.usp_sec_get_filing_metadata @adsh = '0000950170-23-035122';
*/