feat: Add mimic-iv-note (v2.2) loading of postgres

Author: DopamineLcy

mimic-iv-note/buildmimic/postgres/README.md

@@ -0,0 +1,18 @@
+# Load MIMIC-IV-Note into a PostgreSQL database
+
+The scripts in this folder create the schema for MIMIC-IV-Note and load the data into the appropriate tables for PostgreSQL v10+.
+
+## Quickstart
+
+```sh
+# clone repo
+git clone https://github.com/MIT-LCP/mimic-code.git
+cd mimic-code
+# download data
+wget -r -N -c -np --user <USERNAME> --ask-password https://physionet.org/files/mimiciv/2.2/
+mv physionet.org/files/mimiciv mimiciv && rmdir physionet.org/files && rm physionet.org/robots.txt && rmdir physionet.org
+# if mimiciv not exists
+# createdb mimiciv
+psql -d mimiciv -f mimic-iv-note/buildmimic/postgres/create.sql
+psql -d mimiciv -v ON_ERROR_STOP=1 -v mimic_data_dir=mimiciv/2.2 -f mimic-iv-note/buildmimic/postgres/load_gz.sql
+```

mimic-iv-note/buildmimic/postgres/create.sql

@@ -0,0 +1,63 @@
+-------------------------------------------
+-- Create the tables and MIMIC-IV-Note schema --
+-------------------------------------------
+
+----------------------
+-- Creating schemas --
+----------------------
+
+DROP SCHEMA IF EXISTS mimiciv_note CASCADE;
+CREATE SCHEMA mimiciv_note;
+
+---------------------
+-- Creating tables --
+---------------------
+
+-- schema
+
+DROP TABLE IF EXISTS mimiciv_note.discharge;
+CREATE TABLE mimiciv_note.discharge
+(
+  note_id VARCHAR(25) NOT NULL,
+  subject_id INTEGER NOT NULL,
+  hadm_id INTEGER NOT NULL,
+  note_type CHAR(2) NOT NULL,
+  note_seq INTEGER NOT NULL,
+  charttime TIMESTAMP NOT NULL,
+  storetime TIMESTAMP,
+  text TEXT NOT NULL
+);
+
+DROP TABLE IF EXISTS mimiciv_note.radiology;
+CREATE TABLE mimiciv_note.radiology
+(
+  note_id VARCHAR(25) NOT NULL,
+  subject_id INTEGER NOT NULL,
+  hadm_id INTEGER,
+  note_type CHAR(2) NOT NULL,
+  note_seq INTEGER NOT NULL,
+  charttime TIMESTAMP NOT NULL,
+  storetime TIMESTAMP,
+  text TEXT NOT NULL
+);
+
+DROP TABLE IF EXISTS mimiciv_note.discharge_detail;
+CREATE TABLE mimiciv_note.discharge_detail
+(
+  note_id VARCHAR(25) NOT NULL,
+  subject_id INTEGER NOT NULL,
+  field_name VARCHAR(255) NOT NULL,
+  field_value TEXT NOT NULL,
+  field_ordinal INTEGER NOT NULL
+);
+
+
+DROP TABLE IF EXISTS mimiciv_note.radiology_detail;
+CREATE TABLE mimiciv_note.radiology_detail
+(
+  note_id VARCHAR(25) NOT NULL,
+  subject_id INTEGER NOT NULL,
+  field_name VARCHAR(255) NOT NULL,
+  field_value TEXT NOT NULL,
+  field_ordinal INTEGER NOT NULL
+);

mimic-iv-note/buildmimic/postgres/load_gz.sql

@@ -0,0 +1,15 @@
+-----------------------------------------
+-- Load data into the MIMIC-IV-Note schemas --
+-----------------------------------------
+
+-- To run from a terminal:
+--  psql "dbname=<DBNAME> user=<USER>" -v mimic_data_dir=<PATH TO DATA DIR> -f load_gz.sql
+\cd :mimic_data_dir
+
+-- making sure that all tables are emtpy and correct encoding is defined -utf8- 
+SET CLIENT_ENCODING TO 'utf8';
+
+\COPY mimiciv_note.discharge FROM PROGRAM 'gzip -dc discharge.csv.gz' DELIMITER ',' CSV HEADER NULL '';
+\COPY mimiciv_note.radiology FROM PROGRAM 'gzip -dc radiology.csv.gz' DELIMITER ',' CSV HEADER NULL '';
+\COPY mimiciv_note.discharge_detail FROM PROGRAM 'gzip -dc discharge_detail.csv.gz' DELIMITER ',' CSV HEADER NULL '';
+\COPY mimiciv_note.radiology_detail FROM PROGRAM 'gzip -dc radiology_detail.csv.gz' DELIMITER ',' CSV HEADER NULL '';