Window

FIRST_VALUE

DuckDB

Mengembalikan nilai pertama dalam window frame. Sangat berguna untuk mendapatkan nilai awal dalam setiap partisi.

Tipe hasil: Sama dengan tipe expressionDiperbarui: 6 Jan 2026

Syntax

SQL

FIRST_VALUE(expression) OVER ([PARTITION BY col] ORDER BY col [frame_clause])

Parameter

expressionanywajib

Kolom atau ekspresi yang akan diambil nilai pertamanya

PARTITION BYclauseopsional

Opsional. Membagi hasil menjadi partisi

ORDER BYclauseopsional

Menentukan urutan dalam window

frame_clauseclauseopsional

Opsional. Mendefinisikan window frame

Contoh Penggunaan

Harga Awal per Produk

SQL

1 SELECT
2   product_id,
3   date,
4   price,
5   FIRST_VALUE(price) OVER (
6     PARTITION BY product_id
7     ORDER BY date
8     ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING
9   ) as initial_price
10 FROM price_history;

Mendapatkan harga awal setiap produk.

Hasil

product_id	date	price	initial_price
1	2024-01-01	100000	100000
1	2024-02-01	120000	100000
1	2024-03-01	115000	100000

Perbandingan dengan Nilai Awal

SQL

1 SELECT
2   date,
3   revenue,
4   FIRST_VALUE(revenue) OVER (ORDER BY date) as first_revenue,
5   revenue - FIRST_VALUE(revenue) OVER (ORDER BY date) as growth_from_start
6 FROM monthly_revenue;

Membandingkan revenue saat ini dengan revenue awal.

Employee Pertama per Department

SQL

1 SELECT
2   department,
3   name,
4   hire_date,
5   FIRST_VALUE(name) OVER (
6     PARTITION BY department
7     ORDER BY hire_date
8   ) as first_hire
9 FROM employees;

Menampilkan karyawan pertama (senior) di setiap department.

Fungsi Terkait

LAGLEADLAST_VALUENTH_VALUE

Kembali ke DuckDB Functions

1	SELECT
2	product_id,
3	date,
4	price,
5	FIRST_VALUE(price) OVER (
6	PARTITION BY product_id
7	ORDER BY date
8	ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING
9	) as initial_price
10	FROM price_history;

1	SELECT
2	date,
3	revenue,
4	FIRST_VALUE(revenue) OVER (ORDER BY date) as first_revenue,
5	revenue - FIRST_VALUE(revenue) OVER (ORDER BY date) as growth_from_start
6	FROM monthly_revenue;

1	SELECT
2	department,
3	name,
4	hire_date,
5	FIRST_VALUE(name) OVER (
6	PARTITION BY department
7	ORDER BY hire_date
8	) as first_hire
9	FROM employees;