CloudBlobContainer.ListBlobs() - using ToList() to reduce transaction cost
I have two code examples :
var container = new CloudBlobContainer("address");
va开发者_开发问答r blobs = container.ListBlobs();
foreach (var blob in blobs)
{
//Do somthing
}
and this :
var container = new CloudBlobContainer("address");
var blobs = container.ListBlobs().ToList();
foreach (var blob in blobs)
{
//Do somthing
}
Will the second example give any advantage "transaction-wise" ?
How many transaction are made to the blob storage in each example ?Edit: It appears this question is in reference to a blog post by Martin Ingvar Kofoed Jensen. The difference between this question and that question is the call to .Where(...).SingleOrDefault()
. Because LINQ is lazy-evaluated, at the stage of Container.ListBlobs()
it is still an IEnumerable
and hasn't yet called the REST API (no data has been retrieved). As soon as a result operation occurs (such as ToList()
or SingleOrDefault()
), the data is downloaded. Since a non-lazy operation is called on a lazy list inside the loop in the following code, it incurs a transaction for every loop iteration:
foreach (string filePath in allFilesInStartFolder)
{
string fileHash = GetFileHashFromCache(filePath, lastSync);
/* Checking for added files */
var blob = blobs.Where(b => b.LocalPath == filePath).SingleOrDefault();
// ^^ This is a non-lazy op on a lazy evalution, so it causes a REST call.
....
}
Regarding the question as it's actually written: Both code snippets will incur a single transaction (up to 5,000 blobs). I tested both code snippets in LinqPad/Fiddler, and I only see a single API call to generate the list of blobs (replaced our storage name with {mystore}):
https://{mystore}.blob.core.windows.net/
{mystore}?restype=container&comp=list&delimiter=%2F&timeout=90
According to the documentation for List Blobs REST API, up to 5,000 results can be returned in a single call. So if you want to list all blobs in the entire container, it will take at most (# blobs / 5000) transactions.
精彩评论