From d8b526a8b57b5005920a18e9cd834bf42f42062e Mon Sep 17 00:00:00 2001
From: "Peter J. Holzer" <hjp@hjp.at>
Date: Sat, 9 Mar 2024 12:08:38 +0100
Subject: [PATCH] Print size statistics for each column of a table

---
 reports/column_stats | 32 ++++++++++++++++++++++++++++++++
 1 file changed, 32 insertions(+)
 create mode 100755 reports/column_stats

diff --git a/reports/column_stats b/reports/column_stats
new file mode 100755
index 0000000..0dca338
--- /dev/null
+++ b/reports/column_stats
@@ -0,0 +1,32 @@
+#!/usr/bin/python3
+import argparse
+import psycopg2
+import psycopg2.extras
+
+ap = argparse.ArgumentParser()
+ap.add_argument("--dbname")
+ap.add_argument("--user")
+ap.add_argument("--host")
+ap.add_argument("--schema", default="public")
+ap.add_argument("table")
+
+args = ap.parse_args()
+
+db = psycopg2.connect(host=args.host, dbname=args.dbname, user=args.user)
+csr = db.cursor(cursor_factory=psycopg2.extras.NamedTupleCursor)
+csr.execute(
+        """
+        select * from information_schema.columns
+        where table_schema = %s and table_name = %s
+        order by ordinal_position
+        """,
+        (args.schema, args.table,))
+for c in csr.fetchall():
+    q = "select" \
+        f" min(pg_column_size({c.column_name}))," \
+        f" avg(pg_column_size({c.column_name}))," \
+        f" max(pg_column_size({c.column_name}))" \
+        f" from {args.schema}.{args.table}"
+    csr.execute(q)
+    r = csr.fetchone()
+    print(f"{c.column_name:7s}: {c.data_type:17s} {r.min} {r.avg:4.1f} {r.max:3d}")