Question

我使用Rails和Postgresql并且有一个非常大的表（1.3M项）称为FeededProducts我在Rails中使用此调用：

@feed_products = FeededProduct.where("category_connection_id IN (?) AND is_available = ?", @page.category_connections.pluck(:id), true)

导致以下SQL调用：

FeededProduct Load (185.8ms)  SELECT "feeded_products".* FROM "feeded_products"  WHERE (category_connection_id IN (8,2579,3263,7791,7838) AND is_available = 't')

我的数据库中有一个索引：

  add_index "feeded_products", ["category_connection_id", "is_available", "store_id"], name: "cc_w_store_index_on_fp", using: :btree
  add_index "feeded_products", ["category_connection_id", "is_available"], name: "cc_index_on_fp", using: :btree

我希望这个SQL比通常的150-300毫秒快得多。这是正常的吗？我在这里遗漏了一些可以改进的东西吗？

编辑：SQL Explain输出（使用eyeballs gem）：

@feed_products = FeededProduct.where("category_connection_id IN (?) AND is_available = ?", @page.category_connections.pluck(:id), true).eyeballs.explain  
   (1.9ms)  SELECT "category_connections"."id" FROM "category_connections"  WHERE "category_connections"."page_id" = $1  [["page_id", 162]]
  FeededProduct Load (208.4ms)  SELECT "feeded_products".* FROM "feeded_products"  WHERE (category_connection_id IN (8,2579,3263,7791,7838) AND is_available = 't')
 => ["
Bitmap Heap Scan on public.feeded_products  (cost=875.39..74405.02 rows=36805 width=1085) (actual time=12.821..44.605 rows=37314 loops=1)
  Output: id, name, description, category, price, currency, in_stock, product_url, image_url, tracking_url, created_at, updated_at, store_id, slug, is_available, price_before, bad_image, category_connection_id
  Recheck Cond: (feeded_products.category_connection_id = ANY ('{8,2579,3263,7791,7838}'::integer[]))
  Filter: feeded_products.is_available
  Heap Blocks: exact=9428
  Buffers: shared hit=9545
  ->  Bitmap Index Scan on cc_index_on_fp  (cost=0.00..866.19 rows=36805 width=0) (actual time=10.617..10.617 rows=37314 loops=1)
        Index Cond: ((feeded_products.category_connection_id = ANY ('{8,2579,3263,7791,7838}'::integer[])) AND (feeded_products.is_available = true))
        Buffers: shared hit=117
Planning time: 0.323 ms
Execution time: 47.059 ms
"]

postgresql中的索引搜索大约需要200毫秒（1.3M项）。这是正常的吗？

0 个答案: