Gather data

tools/hint/dbjsonl.sh

@@ -0,0 +1,38 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# Usage:
+#   ./import_jsonl.sh "postgresql://user:pass@host:5432/dbname" gloss doc /path/to/file.jsonl
+#
+# Notes:
+# - Creates table if it doesn't exist.
+# - Inserts each JSON line into a jsonb column.
+# - Skips blank lines.
+
+DB_URL="${1:?db url}"
+TABLE="${2:?table name}"
+COL="${3:?json column name}"
+FILE="${4:?jsonl file path}"
+
+psql "$DB_URL" -v ON_ERROR_STOP=1 <<SQL
+CREATE TABLE IF NOT EXISTS ${TABLE} (
+  id bigserial PRIMARY KEY,
+  ${COL} jsonb NOT NULL
+);
+SQL
+
+# \copy runs on the client, so we can feed it from a local file.
+# We copy into a 1-column staging table, then cast to jsonb and insert.
+psql "$DB_URL" -v ON_ERROR_STOP=1 <<SQL
+CREATE TEMP TABLE _jsonl_stage(line text);
+
+\\copy _jsonl_stage(line) FROM '${FILE}' WITH (FORMAT text);
+
+INSERT INTO ${TABLE}(${COL})
+SELECT line::jsonb
+FROM _jsonl_stage
+WHERE btrim(line) <> '';
+
+-- optional: show count inserted this run
+SELECT count(*) AS inserted_now FROM _jsonl_stage WHERE btrim(line) <> '';
+SQL