Iterations by indexes

[michail-vestnik]

Hi.

For example:

set("username:victoria", "value")
set("username:metrics", "value")
...

scan("username:*", 10) // last 10

Is it possible to add this much-needed feature?

If not, please advise how to iterate the data without losing high speed and quality Fastcache?

Create a slice (like an index) with bucket indexes or keys or xxhashes, iterate all elements of the slice and Get values?

Thanks

[michail-vestnik]

Ok, i write this solution:

type collection struct {
	Name string
	Keys []string
}

type cache struct {
    buckets   [bucketCount]bucket
    collections []collection
}

c.collections = make([]collection, 0) // in New()

func (c *cache) Collection(name string) {
    coll := &collection{
        Name: name,
        Keys: make([]string, 0),
    }
    c.collections = append(c.collections, *coll)
}

func (c *cache) Set(collection string, k string, v []byte) error {
    var j int
    for index, coll := range c.collections {
        if coll.Name == collection {
            j = index
	}
    }
    c.collections[j].Keys = append(c.collections[j].Keys, k)
    h := xxhash.Sum64([]byte(k))
    idx := h % bucketCount
    return c.buckets[i].Set([]byte(k), v, h)
}

func (c *cache) Group(collection string, count uint) ([]string, error) {
    var (
        g []string
	j int
    )
    for index, coll := range c.collections {
        if coll.Name == collection {
	    j = index
	}
    }
    for _, k := range c.collections[j].Keys[:count] {
	v, e := c.Get(k)
	if e != nil {
		return nil, e
	}
	g = append(g, string(v))
    }
    return g, nil
}

Will this approach slow down the caching a lot?