Normalization hupanga data ili kila ukweli uwe mahali pamoja tu, ikiondoa marudio. Denormalization hurudufu data kwa makusudi ili kufanya reads kuwa haraka zaidi. Schemas nyingi huanza zikiwa normalized na hu-denormalize tu pale ambapo vipimo vinadai hivyo.
Kila jedwali linaelezea entity moja; data inayohusiana inarejelewa kwa foreign key.
-- Normalized: a customer's city is stored once
customers(id, name, city_id)
cities(id, name)
orders(id, customer_id, total)
-- Denormalized: city copied onto every order for fast reporting
orders(id, customer_id, city_name, total)
-- + a summary table maintained by the app or triggers
order_daily_totals(day, city_name, total_sales)
Chagua normalized kama chaguo-msingi kwa usahihi. Denormalize njia mahususi ya read yenye shughuli nyingi tu baada ya EXPLAIN kuonyesha kwamba joins ndizo bottleneck — na upendelee kuweka source of truth ikiwa normalized huku ukidumisha nakala inayotokana nayo (summary table, materialized cache) unayoweza kuijenga upya.
Huu ni uwiano wa msingi wa muundo wa schema: ku-normalize kupita kiasi hufanya dashboards zitembee kwa kusuasua, ku-denormalize kupita kiasi huzalisha data isiyofanana ambayo ni ghali kuioanisha. Kujua kwa nini ungerudufu data — na jinsi ya kuweka nakala iaminike — hutenganisha schema inayoongezeka (scale) na ile inayooza.
Maktaba ya maswali ya mahojiano ya IT yenye majibu ya kina — kutoka Junior hadi Senior.
Changia