Aggregate

APPROX_QUANTILES

BigQuery

Menghitung perkiraan quantile boundaries dari data numerik. Menggunakan algoritma approximate yang efisien untuk dataset besar, mengembalikan array dengan n+1 elemen.

Tipe hasil: ARRAY<T>Diperbarui: 7 Jan 2026

Syntax

SQL

APPROX_QUANTILES(expression, number)

Parameter

expressionnumericwajib

Kolom atau ekspresi numerik yang akan dihitung quantile-nya

numberINT64wajib

Jumlah quantile yang diinginkan (n akan menghasilkan n+1 boundaries)

Contoh Penggunaan

Menghitung Quartiles

SQL

1 SELECT
2   APPROX_QUANTILES(salary, 4) as salary_quartiles
3 FROM `project.dataset.employees`;

Menghitung min, Q1, median, Q3, max dari gaji.

Hasil

salary_quartiles
[5000000, 8500000, 12000000, 18000000, 45000000]

Menghitung Percentiles

SQL

1 SELECT
2   APPROX_QUANTILES(response_time_ms, 100)[OFFSET(50)] as p50,
3   APPROX_QUANTILES(response_time_ms, 100)[OFFSET(90)] as p90,
4   APPROX_QUANTILES(response_time_ms, 100)[OFFSET(99)] as p99
5 FROM `project.dataset.api_logs`;

Mengambil percentile 50, 90, dan 99 dari response time.

Hasil

p50	p90	p99
120	450	1250

Quartiles per Kategori

SQL

1 SELECT
2   product_category,
3   APPROX_QUANTILES(price, 4) as price_distribution
4 FROM `project.dataset.products`
5 GROUP BY product_category;

Melihat distribusi harga per kategori produk.

Hasil

product_category	price_distribution
Electronics	[99000, 500000, 1500000, 5000000, 25000000]
Fashion	[50000, 150000, 350000, 750000, 5000000]

Median dengan APPROX_QUANTILES

SQL

1 SELECT
2   department,
3   APPROX_QUANTILES(salary, 2)[OFFSET(1)] as median_salary
4 FROM `project.dataset.employees`
5 GROUP BY department
6 ORDER BY median_salary DESC;

Menghitung median gaji per departemen.

Hasil

department	median_salary
Engineering	15500000
Marketing	12000000
Operations	9500000

Fungsi Terkait

APPROX_COUNT_DISTINCTAPPROX_TOP_COUNT

Kembali ke BigQuery Functions

1	SELECT
2	APPROX_QUANTILES(salary, 4) as salary_quartiles
3	FROM `project.dataset.employees`;

1	SELECT
2	APPROX_QUANTILES(response_time_ms, 100)[OFFSET(50)] as p50,
3	APPROX_QUANTILES(response_time_ms, 100)[OFFSET(90)] as p90,
4	APPROX_QUANTILES(response_time_ms, 100)[OFFSET(99)] as p99
5	FROM `project.dataset.api_logs`;

1	SELECT
2	product_category,
3	APPROX_QUANTILES(price, 4) as price_distribution
4	FROM `project.dataset.products`
5	GROUP BY product_category;

1	SELECT
2	department,
3	APPROX_QUANTILES(salary, 2)[OFFSET(1)] as median_salary
4	FROM `project.dataset.employees`
5	GROUP BY department
6	ORDER BY median_salary DESC;