Skip to content

Latest commit

 

History

History
468 lines (365 loc) · 14.1 KB

basic-operations.md

File metadata and controls

468 lines (365 loc) · 14.1 KB

Basic Operations Guide

This guide covers the fundamental operations you'll use with Akavache on a daily basis.

Storing Data

InsertObject - Store data with expiration

// Store data with automatic expiration
var user = new User { Id = 1, Name = "John Doe" };
await CacheDatabase.UserAccount.InsertObject("user_1", user, TimeSpan.FromHours(1));

// Store data without expiration (persists until manually removed)
await CacheDatabase.UserAccount.InsertObject("user_1", user);

// Store with specific expiration time
await CacheDatabase.UserAccount.InsertObject("user_1", user, DateTimeOffset.Now.AddDays(7));

Batch Operations

// Store multiple objects efficiently
var users = new Dictionary<string, User>
{
    ["user_1"] = new User { Id = 1, Name = "John" },
    ["user_2"] = new User { Id = 2, Name = "Jane" }
};

await CacheDatabase.UserAccount.InsertObjects(users, TimeSpan.FromHours(2));

Retrieving Data

GetObject - Retrieve stored data

try
{
    var user = await CacheDatabase.UserAccount.GetObject<User>("user_1");
    Console.WriteLine($"Found user: {user.Name}");
}
catch (KeyNotFoundException)
{
    Console.WriteLine("User not found in cache");
}

GetOrFetchObject - Cache-aside pattern

// Get from cache or fetch from source if not found
var user = await CacheDatabase.UserAccount.GetOrFetchObject("user_1",
    async () => {
        // This function is called only if the key is not in cache
        var response = await httpClient.GetAsync($"https://api.example.com/users/1");
        return await response.Content.ReadFromJsonAsync<User>();
    },
    TimeSpan.FromHours(1)); // Cache for 1 hour

Console.WriteLine($"User: {user.Name}");

Deep Dive: The Get or Fetch Pattern

The GetOrFetchObject method implements the cache-aside pattern, which is the most common and fundamental pattern for handling cached data that originates from a remote source, such as a web API. The logic is simple yet powerful:

  1. The application requests data from the cache.
  2. If the data exists in the cache and has not expired, it is returned immediately.
  3. If the data does not exist in the cache (a "cache miss"), the application executes a provided function—typically a network call—to fetch the "fresh" data from the source.
  4. This fresh data is then inserted into the cache before being returned to the application. Subsequent requests for the same data will now be served from the cache until it expires.

This pattern provides a perfect balance of performance (serving data from a local cache) and data freshness (fetching from the source when necessary).

Basic Get-or-Fetch

This example demonstrates the simplest use case: retrieve a user profile from the cache, or fetch it from an API if it's not present.

// The fetch function (the lambda) is only executed if "user_profile" is not in the cache.
var userData = await CacheDatabase.LocalMachine.GetOrFetchObject("user_profile",
    async () => await apiClient.GetUserProfile(userId));

Get-or-Fetch with Expiration

For data that changes over time, it is crucial to set an expiration policy. This example fetches weather data and caches it for 30 minutes. After 30 minutes, the next request will trigger a fresh fetch from the weather API.

var weatherData = await CacheDatabase.LocalMachine.GetOrFetchObject("weather",
    async () => await weatherApi.GetCurrentWeather(),
    DateTimeOffset.Now.AddMinutes(30));

Get-or-Fetch with a Custom Fetch Observable

The fetch function is not limited to async methods; it can return any IObservable. This allows for more complex or reactive data-fetching scenarios.

// This example fetches a value from a reactive stream.
var liveData = await CacheDatabase.LocalMachine.GetOrFetchObject("live_data",
    () => Observable.Interval(TimeSpan.FromSeconds(5))
                  .Select(_ => DateTime.Now.ToString()));

Advanced Pattern: The GetOrFetchObject pattern is ideal for when you need to ensure you have data before proceeding. For scenarios where you want to display stale (cached) data to the user immediately while a fresh version is fetched in the background, see our detailed guide on the Get and Fetch Latest Pattern.

TryGetObject - Safe retrieval without exceptions

var result = await CacheDatabase.UserAccount.TryGetObject<User>("user_1");
if (result.HasValue)
{
    Console.WriteLine($"Found user: {result.Value.Name}");
}
else
{
    Console.WriteLine("User not found in cache");
}

Batch Retrieval

// Get multiple objects at once
var userKeys = new[] { "user_1", "user_2", "user_3" };
var users = await CacheDatabase.UserAccount.GetObjects<User>(userKeys);

foreach (var kvp in users)
{
    Console.WriteLine($"Key: {kvp.Key}, User: {kvp.Value.Name}");
}

Error Handling

Basic Error Handling

public async Task<User> GetUserSafely(int userId)
{
    try
    {
        return await CacheDatabase.UserAccount.GetObject<User>($"user_{userId}");
    }
    catch (KeyNotFoundException)
    {
        // Key doesn't exist - fetch from API
        return await FetchUserFromApi(userId);
    }
    catch (Exception ex)
    {
        // Other errors (serialization, storage, etc.)
        Console.WriteLine($"Cache error: {ex.Message}");
        return await FetchUserFromApi(userId);
    }
}

Reactive Error Handling

// Using reactive extensions for error handling
var user = await CacheDatabase.UserAccount
    .GetObject<User>("user_1")
    .Catch<User, KeyNotFoundException>(ex => 
        Observable.FromAsync(() => FetchUserFromApi(1)))
    .FirstOrDefaultAsync();

Resilient Cache Operations

public class ResilientCacheService
{
    private readonly IBlobCache _cache = CacheDatabase.UserAccount;
    
    public async Task<T> GetWithFallback<T>(string key, Func<Task<T>> fallback, TimeSpan? expiry = null)
    {
        try
        {
            return await _cache.GetObject<T>(key);
        }
        catch (KeyNotFoundException)
        {
            var data = await fallback();
            if (expiry.HasValue)
            {
                await _cache.InsertObject(key, data, expiry.Value);
            }
            return data;
        }
    }
}

Removing Data

Individual Key Deletion

// Remove a specific key
await CacheDatabase.UserAccount.Invalidate("user_1");

// Remove a specific typed object (recommended for better performance)
await CacheDatabase.UserAccount.InvalidateObject<User>("user_1");

Bulk Deletion

// Remove multiple keys at once
var keysToDelete = new[] { "user_1", "user_2", "user_3" };
await CacheDatabase.UserAccount.Invalidate(keysToDelete);

// Remove multiple typed objects
await CacheDatabase.UserAccount.InvalidateObjects<User>(keysToDelete);

Pattern-Based Deletion

// Find and delete keys matching a pattern
var allKeys = await CacheDatabase.UserAccount.GetAllKeys().FirstOrDefaultAsync();
var userKeys = allKeys.Where(key => key.StartsWith("user_")).ToArray();

if (userKeys.Any())
{
    await CacheDatabase.UserAccount.InvalidateObjects<User>(userKeys);
}

Clear All Data

// Clear all data from a cache
await CacheDatabase.UserAccount.InvalidateAll();

// Clear all objects of a specific type
await CacheDatabase.UserAccount.InvalidateAllObjects<User>();

Updating Expiration

Extend Cache Lifetime

// Update expiration without rewriting data (V11.1+ feature)
await CacheDatabase.UserAccount.UpdateExpiration("user_1", TimeSpan.FromDays(7));

// Update expiration for multiple keys
var keys = new[] { "user_1", "user_2", "user_3" };
await CacheDatabase.UserAccount.UpdateExpiration(keys, TimeSpan.FromDays(7));

Conditional Expiration Updates

// Update expiration only if key exists
try
{
    await CacheDatabase.UserAccount.UpdateExpiration("user_1", TimeSpan.FromDays(7));
    Console.WriteLine("Expiration updated successfully");
}
catch (KeyNotFoundException)
{
    Console.WriteLine("Key not found - cannot update expiration");
}

Performance Benefits

The UpdateExpiration method is highly optimized:

  • High Performance: Only updates metadata, leaves cached data untouched
  • SQL Efficiency: Uses targeted UPDATE statements rather than full record replacement
  • Bulk Operations: Update multiple entries in a single transaction
  • No Data Transfer: Avoids expensive serialization/deserialization cycles (up to 250x faster)

Updating Expiration

// Extend expiration for a single cache entry
await CacheDatabase.LocalMachine.UpdateExpiration("api_data", DateTimeOffset.Now.AddHours(2));

// Extend expiration using relative time
await CacheDatabase.LocalMachine.UpdateExpiration("user_session", TimeSpan.FromMinutes(30));

// Update expiration for multiple entries
var keys = new[] { "cache_key1", "cache_key2", "cache_key3" };
await CacheDatabase.LocalMachine.UpdateExpiration(keys, DateTimeOffset.Now.AddDays(1));

Advanced Operations

Conditional Storage

// Store only if key doesn't exist
public async Task<bool> TryInsertIfNotExists<T>(string key, T value, TimeSpan expiry)
{
    try
    {
        await CacheDatabase.UserAccount.GetObject<T>(key);
        return false; // Key already exists
    }
    catch (KeyNotFoundException)
    {
        await CacheDatabase.UserAccount.InsertObject(key, value, expiry);
        return true; // Successfully inserted
    }
}

Atomic Updates

// Update cache atomically
public async Task<T> UpdateObject<T>(string key, Func<T, T> updateFunc, TimeSpan? expiry = null)
{
    var existing = await CacheDatabase.UserAccount.GetObject<T>(key);
    var updated = updateFunc(existing);
    
    if (expiry.HasValue)
    {
        await CacheDatabase.UserAccount.InsertObject(key, updated, expiry.Value);
    }
    else
    {
        await CacheDatabase.UserAccount.InsertObject(key, updated);
    }
    
    return updated;
}

Cache Statistics

// Get cache information
var allKeys = await CacheDatabase.UserAccount.GetAllKeys().FirstOrDefaultAsync();
Console.WriteLine($"Total cached items: {allKeys.Count()}");

// Get cache size (approximate)
var cacheStats = new
{
    TotalKeys = allKeys.Count(),
    UserKeys = allKeys.Count(k => k.StartsWith("user_")),
    ApiKeys = allKeys.Count(k => k.StartsWith("api_")),
    TempKeys = allKeys.Count(k => k.StartsWith("temp_"))
};

Key Naming Best Practices

Consistent Naming Conventions

// Use consistent, hierarchical naming
await cache.InsertObject("user:profile:123", userProfile);
await cache.InsertObject("user:settings:123", userSettings);
await cache.InsertObject("api:products:page:1", products);
await cache.InsertObject("temp:calculation:hash123", result);

Type-Safe Key Generation

public static class CacheKeys
{
    public static string UserProfile(int userId) => $"user:profile:{userId}";
    public static string UserSettings(int userId) => $"user:settings:{userId}";
    public static string ApiResponse(string endpoint, int page) => $"api:{endpoint}:page:{page}";
    public static string TempData(string identifier) => $"temp:data:{identifier}";
}

// Usage
await cache.InsertObject(CacheKeys.UserProfile(123), userProfile);
var profile = await cache.GetObject<UserProfile>(CacheKeys.UserProfile(123));

Common Patterns

Cache-Aside Pattern

public class UserService
{
    private readonly IBlobCache _cache = CacheDatabase.UserAccount;
    private readonly HttpClient _httpClient;
    
    public async Task<User> GetUser(int userId)
    {
        var cacheKey = CacheKeys.UserProfile(userId);
        
        return await _cache.GetOrFetchObject(cacheKey,
            async () => {
                var response = await _httpClient.GetAsync($"/api/users/{userId}");
                return await response.Content.ReadFromJsonAsync<User>();
            },
            TimeSpan.FromMinutes(30));
    }
}

Write-Through Pattern

public class UserService
{
    private readonly IBlobCache _cache = CacheDatabase.UserAccount;
    private readonly HttpClient _httpClient;
    
    public async Task<User> UpdateUser(int userId, User updatedUser)
    {
        // Update the backend
        var response = await _httpClient.PutAsJsonAsync($"/api/users/{userId}", updatedUser);
        var savedUser = await response.Content.ReadFromJsonAsync<User>();
        
        // Update the cache
        var cacheKey = CacheKeys.UserProfile(userId);
        await _cache.InsertObject(cacheKey, savedUser, TimeSpan.FromMinutes(30));
        
        return savedUser;
    }
}

Write-Behind Pattern

public class UserService
{
    private readonly IBlobCache _cache = CacheDatabase.UserAccount;
    private readonly Queue<PendingUpdate> _pendingUpdates = new();
    
    public async Task UpdateUserAsync(int userId, User updatedUser)
    {
        // Update cache immediately
        var cacheKey = CacheKeys.UserProfile(userId);
        await _cache.InsertObject(cacheKey, updatedUser, TimeSpan.FromMinutes(30));
        
        // Queue backend update for later
        _pendingUpdates.Enqueue(new PendingUpdate { UserId = userId, User = updatedUser });
        
        // Process queue asynchronously (implement background service)
        _ = Task.Run(ProcessPendingUpdates);
    }
}

Performance Tips

  1. Use batch operations when working with multiple keys
  2. Choose appropriate expiration times - not too short (cache misses) or too long (stale data)
  3. Use UpdateExpiration instead of re-inserting data when you only need to extend lifetime
  4. Prefer typed operations (InvalidateObject<T>) over untyped ones when possible
  5. Use consistent key naming for easier maintenance and debugging
  6. Handle exceptions gracefully to prevent cache issues from breaking your application
  7. Monitor cache hit/miss ratios to optimize your caching strategy

Next Steps