Struktur data Redis canggih apa lan modul apa sing ana?

Question

Accepted Answer

Lumabar tipe intinya, Redis nggawe struktur khusus — **HyperLogLog**, **bitmaps**, **geospatial**, **bitfields** — lan **modules** (RedisJSON, RediSearch, RedisTimeSeries, RedisBloom) sing ngembangake Redis kanggo kasus pakaryan khusus kanthi efisien.

## HyperLogLog — perkiraan hitungan unik

```bash
# count UNIQUE items using tiny, fixed memory (~12KB) with ~0.81% error
PFADD visitors:today "user1" "user2" "user3"
PFCOUNT visitors:today          # approximate count of unique visitors
PFMERGE visitors:week visitors:mon visitors:tue   # merge HLLs
# → counts millions of uniques in ~12KB (a Set would need huge memory!)
```

**HyperLogLog** nerangake cardinality (hitungan unik) nganggo memori tetep cilik — ideal kanggo ngetung pengunjung unik, IP, utawa acara kanthi skala gedhe dene counting eksak bakal butuh memori abot.

## Bitmaps lan bitfields — gantheng kompak

```bash
# bitmaps: bit-per-item flags (very memory-efficient)
SETBIT active:2024-01-15 1000 1      # mark user 1000 active today
BITCOUNT active:2024-01-15           # how many users were active (count set bits)
BITOP AND active:both d1 d2          # users active BOTH days (bitwise AND)
# → track daily active users for millions in a few KB per day
```

**Bitmaps** nyimpen gantheng per-item kanthi abot kompak — sampurna kanggo tracking daily-active-user, feature flags per pengguna, lan presence kanthi skala.

## Geospatial — query lokasi

```bash
GEOADD places 13.361 38.115 "Palermo"      # add a location (lng, lat)
GEOSEARCH places FROMMEMBER "Palermo" BYRADIUS 200 km ASC   # find nearby
```

## Modul Redis — ngembangake Redis

```text
RedisJSON      → store/query JSON documents natively (paths, partial updates)
RediSearch     → full-text search and secondary indexing on Redis data
RedisTimeSeries→ time-series data (metrics, IoT) with downsampling/aggregation
RedisBloom     → Bloom filters / Cuckoo filters (probabilistic membership)
RedisGraph (legacy), and others
→ Modules turn Redis into a multi-model platform (Redis Stack bundles popular ones).
```

## Apa pule ingone

Ngerteni struktur data canggih Redis lan modul-modul duwe kauntungan pengetahuan tingkat senior kanggo **mbaleni masalah khusus kanthi efisien** nganggo Redis, ngembangake apa sing bisa-bisa Redis luwih saka tipe intinya.

Strutur **probabilistik/kompak** utamane kuat kanggo skala: **HyperLogLog** (nerangake hitungan unik ing memori tetep cilik ~12KB, ideal kanggo ngetung pengunjung/IP/acara unik kanthi skala dene counting eksak liwat sets bakal larang memori — menang memori dramatis), **bitmaps** (gantheng per-item nyimpen babar pisan kaya daily-active-users kanggo jutaan ing sawetara KB, karo operasi bitwise efisien kanggo analitik), lan **bitfields** — iki mbaleni masalah skala bener-bener (cardinality, presence tracking, gantheng) karo efisiensi memori abot sing angel ditandingi cara liya. **Geospatial** mendung nglakokake query lokasi langsung ing Redis.

**Modul-modul** — RedisJSON (dokumen JSON asli), RediSearch (full-text search lan indeks sekunder), RedisTimeSeries (time-series/metrics karo agregasi), RedisBloom (filter keanggotaan probabilistik) — ngembangake Redis dadi **platform multi-model** (bundled ing Redis Stack), ngbolake kanthi nangani search, JSON, time-series, lan kasus probabilistik sing cara liya butuh sistem terpisah.

Ngerteni iki ngbolake leverage Redis kanggo kabutuhan khusus kanthi efisien tinimbang njaluk infrastruktur tambahan utawa terima memori/performance kurang apik.

Kabukti masalah khusus (counting unik skala besar, presence tracking, search, time-series, JSON) muncul ing sistem nyata, lan kabukti struktur canggih Redis (HyperLogLog, bitmaps, geospatial) lan modul-modul (RedisJSON, RediSearch, RedisTimeSeries, RedisBloom) mbaleni kanthi efisien, ngerteni struktur data canggih Redis lan modul duwe kauntungan pengetahuan tingkat senior sing nampakake kemampuan Redis lengkap minangka platform versatile multi-model, ngebole solusi elegan lan hemat memori kanggo masalah khusus lan cetha keahlian Redis jero tekan luwih ana saka kasus caching umum.