You cannot select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
130 lines
3.4 KiB
SQL
130 lines
3.4 KiB
SQL
CREATE SCHEMA http;
|
|
/*
|
|
The purpose of this schema, tables, and associated roles to process HTTP responses.
|
|
I may even include a table to keep track of the XML responses
|
|
*/
|
|
|
|
/*
|
|
Add a role to manage permissions on the http schema
|
|
*/
|
|
CREATE ROLE http_requestor;
|
|
|
|
GRANT CONNECT ON DATABASE aact_db to http_requestor;
|
|
|
|
GRANT USAGE ON SCHEMA http TO http_requestor;
|
|
|
|
GRANT INSERT,SELECT ON ALL TABLES IN SCHEMA http TO http_requestor;
|
|
|
|
|
|
/* Create tables related to http requests
|
|
As not every request will have an xml doc, split them.
|
|
*/
|
|
|
|
CREATE TABLE IF NOT EXISTS http.responses (
|
|
id SERIAL PRIMARY KEY,
|
|
nct_id VARCHAR(15),
|
|
version_a SMALLINT,
|
|
version_b SMALLINT,
|
|
url VARCHAR(255),
|
|
response_code SMALLINT,
|
|
response_date TIMESTAMP WITH TIME ZONE,
|
|
html TEXT
|
|
);
|
|
|
|
|
|
|
|
|
|
-- Type: history_download_status
|
|
-- DROP TYPE IF EXISTS http.history_download_status;
|
|
|
|
CREATE TYPE http.history_download_status AS ENUM
|
|
('Ignored', 'Of Interest', 'Reserved', 'Downloaded', 'Incomplete');
|
|
|
|
ALTER TYPE http.history_download_status
|
|
OWNER TO root;
|
|
|
|
COMMENT ON TYPE http.history_download_status
|
|
IS 'Records the status of the downloaded history from clinicaltrials.gov';
|
|
|
|
/*
|
|
Table: http.download_status
|
|
|
|
Creates and then initializes the table to record history download status.
|
|
*/
|
|
|
|
--DROP TABLE IF EXISTS http.download_status;
|
|
|
|
CREATE TABLE IF NOT EXISTS http.download_status
|
|
(
|
|
id SERIAL NOT NULL,
|
|
nct_id character varying(15) COLLATE pg_catalog."default" NOT NULL,
|
|
status http.history_download_status NOT NULL DEFAULT 'Ignored'::http.history_download_status,
|
|
update_timestamp timestamp with time zone DEFAULT now(),
|
|
CONSTRAINT download_status_pkey PRIMARY KEY (id),
|
|
CONSTRAINT "nct_id link" FOREIGN KEY (nct_id)
|
|
REFERENCES ctgov.studies (nct_id) MATCH SIMPLE
|
|
ON UPDATE NO ACTION
|
|
ON DELETE NO ACTION
|
|
)
|
|
|
|
TABLESPACE pg_default;
|
|
|
|
ALTER TABLE IF EXISTS http.download_status
|
|
OWNER to root;
|
|
|
|
COMMENT ON TABLE http.download_status
|
|
IS 'Records the download status for each trial''s history';
|
|
|
|
|
|
INSERT INTO http.download_status (nct_id, status)
|
|
SELECT DISTINCT nct_id, 'Ignored'::http.history_download_status
|
|
FROM ctgov.studies;
|
|
|
|
|
|
|
|
/*
|
|
Views to allow downloader programs to identify what to download.
|
|
*/
|
|
|
|
-- View: http.most_recent_download_status
|
|
|
|
-- DROP VIEW http.most_recent_download_status;
|
|
|
|
CREATE OR REPLACE VIEW http.most_recent_download_status
|
|
AS
|
|
SELECT t.nct_id,
|
|
t.status,
|
|
t.update_timestamp
|
|
FROM ( SELECT download_status.id,
|
|
download_status.nct_id,
|
|
download_status.status,
|
|
download_status.update_timestamp,
|
|
row_number() OVER (PARTITION BY download_status.nct_id ORDER BY download_status.update_timestamp DESC) AS rn
|
|
FROM http.download_status) t
|
|
WHERE t.rn = 1
|
|
ORDER BY t.nct_id;
|
|
|
|
ALTER TABLE http.most_recent_download_status
|
|
OWNER TO root;
|
|
COMMENT ON VIEW http.most_recent_download_status
|
|
IS 'List of most recent download status';
|
|
|
|
|
|
|
|
-- View: http.trials_to_download
|
|
|
|
-- DROP VIEW http.trials_to_download;
|
|
|
|
CREATE OR REPLACE VIEW http.trials_to_download
|
|
AS
|
|
SELECT most_recent_download_status.nct_id
|
|
FROM http.most_recent_download_status
|
|
WHERE most_recent_download_status.status = 'Of Interest'::http.history_download_status;
|
|
|
|
ALTER TABLE http.trials_to_download
|
|
OWNER TO root;
|
|
COMMENT ON VIEW http.trials_to_download
|
|
IS 'List of trials for which the history needs downloaded';
|
|
|
|
|