Azure-sdk-for-python: list_blobs only list 5000 elements

[jmlero]

Hi all,

I am trying to download a "folder" inside of a blob container keeping the original tree folder structure. This "folder" contains millions of files.
For this reason first I run list_blobs just to obtain the list of blobs and then download the blobs using get_blob_to_path.

blobs = blob_service.list_blobs('blob_container','data/projects/folder')
for blob in blobs:
print(blob.name)

this function only shows the first 5000 blobs inside of data/projects/folder, but as I said, I have millions of files.

Any idea why this loop only shows the first 5000 elements?
Any other suggestion to download millions of files inside of a blob container?

Thanks in advance and best regards?

[Segsfault]

You'll need to set a marker for containers with > 5k entries. here is a code snippet I used:

marker = None
while True:
  results = blob_service.list_blobs('blob_container',marker=marker, prefix=prefix, delimiter=delimiter)
  #...do stuff with results ...
  if results.next_marker:
    marker = results.next_marker
  else:
    break

Basically you set the initial marker to None to start at the beginning, then loop until a result set does not return a pointer to a new marker.

[jmlero]

Hi,

Really, thanks a lot, is working now. Altough I am not using "prefix" and "delimiter" since I am not sure what is the purpose of those two parameters.

Best regards,

[Segsfault]

Prefix and delimiter are utility parameters for fetching file lists. For example let's say you have a directory structure like this:

.
├── data
│   ├── archive
│   │   ├── old_file1.csv
│   │   └── old_file2.csv
│   ├── file1.csv
│   ├── file2.csv
│   └── file3.csv
└── logs
    └── log.txt

If you set the prefix to 'data', you will get:

data/file1.csv
data/file2.csv
data/file3.csv
data/archive/old_file1.csv
data/archive/old_file2.csv

But if you also set the delimiter to '/', you will only get:

data/file1.csv
data/file2.csv
data/file3.csv

In this example, I used prefix to target a specific directory, and a delimiter of '/' to specify that I only want the files in that directory (and not subdirectories and their contents below it).

Many applications of these parameters, but this is what I use them for anyways.