Struct opendal::Operator

pub struct Operator { /* private fields */ }

Operator is the entry for all public async APIs.

Developer should manipulate the data from storage service through Operator only by right.

We will usually do some general checks and data transformations in this layer, like normalizing path from input, checking whether the path refers to one file or one directory, and so on. Read Operator::concepts for more about [Operator::Operator].

Examples

Read more backend init examples in [Operator::services]

use opendal::services::Fs;
use opendal::Operator;
async fn test() -> Result<()> {
    // Create fs backend builder.
    let mut builder = Fs::default();
    // Set the root for fs, all operations will happen under this root.
    //
    // NOTE: the root must be absolute path.
    builder.root("/tmp");

    // Build an `Operator` to start operating the storage.
    let _: Operator = Operator::new(builder)?.finish();

    Ok(())
}

Implementations

impl Operator

Operator basic API.

pub fn limit(&self) -> usize

Get current operator’s limit. Limit is usually the maximum size of data that operator will handle in one operation.

pub fn with_limit(&self, limit: usize) -> Self

Specify the batch limit.

Default: 1000

pub fn info(&self) -> OperatorInfo

Get information of underlying accessor.

Examples

use opendal::Operator;

let info = op.info();

pub fn blocking(&self) -> BlockingOperator

Create a new blocking operator.

This operation is nearly no cost.

impl Operator

Operator async API.

pub async fn check(&self) -> Result<()>

Check if this operator can work correctly.

We will send a list request to path and return any errors we met.

use opendal::Operator;

op.check().await?;

pub async fn stat(&self, path: &str) -> Result<Metadata>

Get given path’s metadata.

Notes

Extra Options

Operator::stat is a wrapper of Operator::stat_with without any options. To use extra options like if_match and if_none_match, please use Operator::stat_with instead.

Reuse Metadata

For fetch metadata of entries returned by [Operator::Lister], it’s better to use Operator::list_with and Operator::lister_with with metakey query like Metakey::ContentLength | Metakey::LastModified so that we can avoid extra stat requests.

Examples

Check if file exists

use opendal::ErrorKind;
if let Err(e) = op.stat("test").await {
    if e.kind() == ErrorKind::NotFound {
        println!("file not exist")
    }
}

pub fn stat_with( &self, path: &str ) -> FutureStat<impl Future<Output = Result<Metadata>>>

Get given path’s metadata with extra options.

Notes

Reuse Metadata

For fetch metadata of entries returned by [Operator::Lister], it’s better to use Operator::list_with and Operator::lister_with with metakey query like Metakey::ContentLength | Metakey::LastModified so that we can avoid extra requests.

Options

if_match

Set if_match for this stat request.

This feature can be used to check if the file’s ETag matches the given ETag.

If file exists and it’s etag doesn’t match, an error with kind ErrorKind::ConditionNotMatch will be returned.

use opendal::Operator;

let mut metadata = op.stat_with("path/to/file").if_match(etag).await?;

if_none_match

Set if_none_match for this stat request.

This feature can be used to check if the file’s ETag doesn’t match the given ETag.

If file exists and it’s etag match, an error with kind ErrorKind::ConditionNotMatch will be returned.

use opendal::Operator;

let mut metadata = op.stat_with("path/to/file").if_none_match(etag).await?;

Examples

Get metadata while ETag matches

stat_with will

• return Ok(metadata) if ETag matches
• return Err(error) and error.kind() == ErrorKind::ConditionNotMatch if file exists but ETag mismatch
• return Err(err) if other errors occur, for example, NotFound.

use opendal::ErrorKind;
if let Err(e) = op.stat_with("test").if_match("<etag>").await {
    if e.kind() == ErrorKind::ConditionNotMatch {
        println!("file exists, but etag mismatch")
    }
    if e.kind() == ErrorKind::NotFound {
        println!("file not exist")
    }
}

---

Behavior

Services that support create_dir

test and test/ may vary in some services such as S3. However, on a local file system, they’re identical. Therefore, the behavior of stat("test") and stat("test/") might differ in certain edge cases. Always use stat("test/") when you need to access a directory if possible.

Here are the behavior list:

Case	Path	Result
stat existing dir	abc/	Metadata with dir mode
stat existing file	abc/def_file	Metadata with file mode
stat dir without /	abc/def_dir	Error NotFound or metadata with dir mode
stat file with /	abc/def_file/	Error NotFound
stat not existing path	xyz	Error NotFound

Refer to RFC: List Prefix for more details.

Services that not support create_dir

For services that not support create_dir, stat("test/") will return NotFound even when test/abc exists since the service won’t have the concept of dir. There is nothing we can do about this.

pub async fn is_exist(&self, path: &str) -> Result<bool>

Check if this path exists or not.

Example

use anyhow::Result;
use futures::io;
use opendal::Operator;

async fn test(op: Operator) -> Result<()> {
    let _ = op.is_exist("test").await?;

    Ok(())
}

pub async fn create_dir(&self, path: &str) -> Result<()>

Create a dir at given path.

Notes

To indicate that a path is a directory, it is compulsory to include a trailing / in the path. Failure to do so may result in NotADirectory error being returned by OpenDAL.

Behavior

• Create on existing dir will succeed.
• Create dir is always recursive, works like mkdir -p

Examples

op.create_dir("path/to/dir/").await?;

pub async fn read(&self, path: &str) -> Result<Buffer>

Read the whole path into a bytes.

Notes

Extra Options

Operator::read is a wrapper of Operator::read_with without any options. To use extra options like range and if_match, please use Operator::read_with instead.

Streaming Read

This function will allocate a new bytes internally. For more precise memory control or reading data lazily, please use Operator::reader

Examples

let bs = op.read("path/to/file").await?;

pub fn read_with( &self, path: &str ) -> FutureRead<impl Future<Output = Result<Buffer>>>

Read the whole path into a bytes with extra options.

This function will allocate a new bytes internally. For more precise memory control or reading data lazily, please use Operator::reader

Notes

Streaming Read

This function will allocate a new bytes internally. For more precise memory control or reading data lazily, please use Operator::reader

Options

range

Set range for this read request.

If we have a file with size n.

• .. means read bytes in range [0, n) of file.
• 0..1024 means read bytes in range [0, 1024) of file
• 1024.. means read bytes in range [1024, n) of file
• ..1024 means read bytes in range (n - 1024, n) of file

let bs = op.read_with("path/to/file").range(0..1024).await?;

if_match

Set if_match for this read request.

This feature can be used to check if the file’s ETag matches the given ETag.

If file exists and it’s etag doesn’t match, an error with kind ErrorKind::ConditionNotMatch will be returned.

use opendal::Operator;
let mut metadata = op.read_with("path/to/file").if_match(etag).await?;

if_none_match

Set if_none_match for this read request.

This feature can be used to check if the file’s ETag doesn’t match the given ETag.

If file exists and it’s etag match, an error with kind ErrorKind::ConditionNotMatch will be returned.

use opendal::Operator;
let mut metadata = op.read_with("path/to/file").if_none_match(etag).await?;

concurrent

Set concurrent for the reader.

OpenDAL by default to write file without concurrent. This is not efficient for cases when users read large chunks of data. By setting concurrent, opendal will read files concurrently on support storage services.

By setting concurrent, opendal will fetch chunks concurrently with the given chunk size.

let r = op.read_with("path/to/file").concurrent(8).await?;

chunk

OpenDAL will use services’ preferred chunk size by default. Users can set chunk based on their own needs.

This following example will make opendal read data in 4MiB chunks:

let r = op.read_with("path/to/file").chunk(4 * 1024 * 1024).await?;

Examples

Read the whole path into a bytes.

let bs = op.read_with("path/to/file").await?;
let bs = op.read_with("path/to/file").range(0..10).await?;

pub async fn reader(&self, path: &str) -> Result<Reader>

Create a new reader which can read the whole path.

Notes

Extra Options

Operator::reader is a wrapper of Operator::reader_with without any options. To use extra options like concurrent, please use Operator::reader_with instead.

Examples

let r = op.reader("path/to/file").await?;

pub fn reader_with( &self, path: &str ) -> FutureReader<impl Future<Output = Result<Reader>>>

Create a new reader with extra options

Notes

Extra Options

Operator::reader is a wrapper of Operator::reader_with without any options. To use extra options like version, please use Operator::reader_with instead.

Options

concurrent

Set concurrent for the reader.

OpenDAL by default to write file without concurrent. This is not efficient for cases when users read large chunks of data. By setting concurrent, opendal will reading files concurrently on support storage services.

By setting `concurrent``, opendal will fetch chunks concurrently with the give chunk size.

let r = op.reader_with("path/to/file").concurrent(8).await?;

chunk

OpenDAL will use services’ preferred chunk size by default. Users can set chunk based on their own needs.

This following example will make opendal read data in 4MiB chunks:

let r = op.reader_with("path/to/file").chunk(4 * 1024 * 1024).await?;

Examples

let r = op.reader_with("path/to/file").version("version_id").await?;

pub async fn write(&self, path: &str, bs: impl Into<Buffer>) -> Result<()>

Write bytes into path.

Notes

Extra Options

Operator::write is a wrapper of Operator::write_with without any options. To use extra options like content_type and cache_control, please use Operator::write_with instead.

Streaming Write

This function will write all bytes at once. For more precise memory control or writing data continuously, please use Operator::writer.

Multipart Uploads Write

OpenDAL abstracts the multipart uploads into Writer. It will automatically handle the multipart uploads for you. You can control the behavior of multipart uploads by setting chunk, concurrent via Operator::writer_with

Examples

use bytes::Bytes;

op.write("path/to/file", vec![0; 4096]).await?;

pub async fn copy(&self, from: &str, to: &str) -> Result<()>

Copy a file from from to to.

Notes

• from and to must be a file.
• to will be overwritten if it exists.
• If from and to are the same, an IsSameFile error will occur.
• copy is idempotent. For same from and to input, the result will be the same.

Examples

op.copy("path/to/file", "path/to/file2").await?;

pub async fn rename(&self, from: &str, to: &str) -> Result<()>

Rename a file from from to to.

Notes

• from and to must be a file.
• to will be overwritten if it exists.
• If from and to are the same, an IsSameFile error will occur.

Examples

op.rename("path/to/file", "path/to/file2").await?;

pub async fn writer(&self, path: &str) -> Result<Writer>

Write multiple bytes into path.

Notes

Extra Options

Operator::write is a wrapper of Operator::write_with without any options. To use extra options like content_type and cache_control, please use Operator::write_with instead.

Chunk

OpenDAL is designed to write files directly without chunking by default, giving users control over the exact size of their writes and helping avoid unnecessary costs.

This is not efficient for cases when users write small chunks of data. Some storage services like s3 could even return hard errors like EntityTooSmall. Besides, cloud storage services will cost more money if we write data in small chunks.

Users can use Operator::write_with to set a good chunk size might improve the performance,

Examples

use bytes::Bytes;

let mut w = op.writer("path/to/file").await?;
w.write(vec![0; 4096]).await?;
w.write(vec![1; 4096]).await?;
w.close().await?;

pub fn writer_with( &self, path: &str ) -> FutureWriter<impl Future<Output = Result<Writer>>>

Write multiple bytes into path with extra options.

Options

append

Set append for this write request.

write by default to overwrite existing files. To append to the end of file instead, please set append to true.

The following example will append data to existing file instead.

use bytes::Bytes;

let mut w = op.writer_with("path/to/file").append(true).await?;
w.write(vec![0; 4096]).await?;
w.write(vec![1; 4096]).await?;
w.close().await?;

chunk

Set chunk for the writer.

OpenDAL is designed to write files directly without chunking by default, giving users control over the exact size of their writes and helping avoid unnecessary costs.

This is not efficient for cases when users write small chunks of data. Some storage services like s3 could even return hard errors like EntityTooSmall. Besides, cloud storage services will cost more money if we write data in small chunks.

Set a good chunk size might improve the performance, reduce the API calls and save money.

The following example will set the writer chunk to 8MiB. Only one API call will be sent at close instead.

use bytes::Bytes;

let mut w = op
    .writer_with("path/to/file")
    .chunk(8 * 1024 * 1024)
    .await?;
w.write(vec![0; 4096]).await?;
w.write(vec![1; 4096]).await?;
w.close().await?;

concurrent

Set concurrent for the writer.

OpenDAL by default to write file without concurrent. This is not efficient for cases when users write large chunks of data. By setting concurrent, opendal will writing files concurrently on support storage services.

The following example will set the writer concurrent to 8.

• The first write will start and return immediately.
• The second write will start and return immediately.
• The close will make sure all writes are done in order and return result.

use bytes::Bytes;

let mut w = op.writer_with("path/to/file").concurrent(8).await?;
w.write(vec![0; 4096]).await?; // Start the first write
w.write(vec![1; 4096]).await?; // Second write will be concurrent without wait
w.close().await?; // Close will make sure all writes are done and success

cache_control

Set the cache_control for this write request.

Some storage services support setting cache_control as system metadata.

use bytes::Bytes;

let mut w = op
    .writer_with("path/to/file")
    .cache_control("max-age=604800")
    .await?;
w.write(vec![0; 4096]).await?;
w.write(vec![1; 4096]).await?;
w.close().await?;

content_type

Set the content_type for this write request.

Some storage services support setting content_type as system metadata.

use bytes::Bytes;

let mut w = op
    .writer_with("path/to/file")
    .content_type("text/plain")
    .await?;
w.write(vec![0; 4096]).await?;
w.write(vec![1; 4096]).await?;
w.close().await?;

content_disposition

Set the content_disposition for this write request.

Some storage services support setting content_disposition as system metadata.

use bytes::Bytes;

let mut w = op
    .writer_with("path/to/file")
    .content_disposition("attachment; filename=\"filename.jpg\"")
    .await?;
w.write(vec![0; 4096]).await?;
w.write(vec![1; 4096]).await?;
w.close().await?;

Examples

use bytes::Bytes;

let mut w = op
    .writer_with("path/to/file")
    .content_type("application/octet-stream")
    .await?;
w.write(vec![0; 4096]).await?;
w.write(vec![1; 4096]).await?;
w.close().await?;

pub fn write_with( &self, path: &str, bs: impl Into<Buffer> ) -> FutureWrite<impl Future<Output = Result<()>>>

Write data with extra options.

Notes

Streaming Write

This function will write all bytes at once. For more precise memory control or writing data lazily, please use Operator::writer_with.

Multipart Uploads Write

OpenDAL abstracts the multipart uploads into Writer. It will automatically handle the multipart uploads for you. You can control the behavior of multipart uploads by setting chunk, concurrent via Operator::writer_with

Options

append

Set append for this write request.

write by default to overwrite existing files. To append to the end of file instead, please set append to true.

The following example will append data to existing file instead.

use bytes::Bytes;

let bs = b"hello, world!".to_vec();
let _ = op.write_with("path/to/file", bs).append(true).await?;

cache_control

Set the cache_control for this write request.

Some storage services support setting cache_control as system metadata.

use bytes::Bytes;

let bs = b"hello, world!".to_vec();
let _ = op
    .write_with("path/to/file", bs)
    .cache_control("max-age=604800")
    .await?;

content_type

Set the content_type for this write request.

Some storage services support setting content_type as system metadata.

use bytes::Bytes;

let bs = b"hello, world!".to_vec();
let _ = op
    .write_with("path/to/file", bs)
    .content_type("text/plain")
    .await?;

content_disposition

Set the content_disposition for this write request.

Some storage services support setting content_disposition as system metadata.

use bytes::Bytes;

let bs = b"hello, world!".to_vec();
let _ = op
    .write_with("path/to/file", bs)
    .content_disposition("attachment; filename=\"filename.jpg\"")
    .await?;

Examples

use bytes::Bytes;

let bs = b"hello, world!".to_vec();
let _ = op
    .write_with("path/to/file", bs)
    .content_type("text/plain")
    .await?;

pub async fn delete(&self, path: &str) -> Result<()>

Delete the given path.

Notes

• Deleting a file that does not exist won’t return errors.

Examples

op.delete("test").await?;

pub fn delete_with( &self, path: &str ) -> FutureDelete<impl Future<Output = Result<()>>>

Delete the given path with extra options.

Notes

• Deleting a file that does not exist won’t return errors.

Examples

op.delete_with("test").await?;

pub async fn remove(&self, paths: Vec<String>) -> Result<()>

Notes

If underlying services support delete in batch, we will use batch delete instead.

Examples

op.remove(vec!["abc".to_string(), "def".to_string()])
    .await?;

pub async fn remove_via( &self, input: impl Stream<Item = String> + Unpin ) -> Result<()>

remove will remove files via the given paths.

remove_via will remove files via the given stream.

We will delete by chunks with given batch limit on the stream.

Notes

If underlying services support delete in batch, we will use batch delete instead.

Examples

use futures::stream;
let stream = stream::iter(vec!["abc".to_string(), "def".to_string()]);
op.remove_via(stream).await?;

pub async fn remove_all(&self, path: &str) -> Result<()>

Remove the path and all nested dirs and files recursively.

Notes

If underlying services support delete in batch, we will use batch delete instead.

Examples

op.remove_all("path/to/dir").await?;

pub async fn list(&self, path: &str) -> Result<Vec<Entry>>

List entries that starts with given path in parent dir.

Notes

Recursively List

This function only read the children of the given directory. To read all entries recursively, use Operator::list_with("path").recursive(true) instead.

Streaming List

This function will read all entries in the given directory. It could take very long time and consume a lot of memory if the directory contains a lot of entries.

In order to avoid this, you can use Operator::lister to list entries in a streaming way.

Reuse Metadata

The only metadata that is guaranteed to be available is the Mode. For fetching more metadata, please use Operator::list_with and metakey.

Examples

List entries under a dir

This example will list all entries under the dir path/to/dir/.

use opendal::EntryMode;
use opendal::Metakey;
use opendal::Operator;
let mut entries = op.list("path/to/dir/").await?;
for entry in entries {
    match entry.metadata().mode() {
        EntryMode::FILE => {
            println!("Handling file")
        }
        EntryMode::DIR => {
            println!("Handling dir {}", entry.path())
        }
        EntryMode::Unknown => continue,
    }
}

List entries with prefix

This example will list all entries under the prefix path/to/prefix.

NOTE: it’s possible that the prefix itself is also a dir. In this case, you could get path/to/prefix/, path/to/prefix_1 and so on. If you do want to list a dir, please make sure the path is end with /.

use opendal::EntryMode;
use opendal::Metakey;
use opendal::Operator;
let mut entries = op.list("path/to/prefix").await?;
for entry in entries {
    match entry.metadata().mode() {
        EntryMode::FILE => {
            println!("Handling file")
        }
        EntryMode::DIR => {
            println!("Handling dir {}", entry.path())
        }
        EntryMode::Unknown => continue,
    }
}

pub fn list_with( &self, path: &str ) -> FutureList<impl Future<Output = Result<Vec<Entry>>>>

List entries that starts with given path in parent dir with more options.

Notes

Streaming list

This function will read all entries in the given directory. It could take very long time and consume a lot of memory if the directory contains a lot of entries.

In order to avoid this, you can use Operator::lister to list entries in a streaming way.

Options

start_after

Specify the specified key to start listing from.

This feature can be used to resume a listing from a previous point.

The following example will resume the list operation from the breakpoint.

use opendal::Operator;
let mut entries = op
    .list_with("path/to/dir/")
    .start_after("breakpoint")
    .await?;

recursive

Specify whether to list recursively or not.

If recursive is set to true, we will list all entries recursively. If not, we’ll only list the entries in the specified dir.

use opendal::Operator;
let mut entries = op.list_with("path/to/dir/").recursive(true).await?;

metakey

Specify the metadata that required to be fetched in entries.

If metakey is not set, we will fetch only the entry’s mode. Otherwise, we will retrieve the required metadata from storage services. Even if metakey is specified, the metadata may still be None, indicating that the storage service does not supply this information.

Some storage services like s3 could return more metadata like content-length and last-modified. By using metakey, we can fetch those metadata without an extra stat call. Please pick up the metadata you need to reduce the extra stat cost.

This example shows how to list entries with content-length and last-modified metadata:

use opendal::EntryMode;
use opendal::Metakey;
use opendal::Operator;
let mut entries = op
    .list_with("dir/")
    // Make sure content-length and last-modified been fetched.
    .metakey(Metakey::ContentLength | Metakey::LastModified)
    .await?;
for entry in entries {
    let meta = entry.metadata();
    match meta.mode() {
        EntryMode::FILE => {
            println!(
                "Handling file {} with size {}",
                entry.path(),
                meta.content_length()
            )
        }
        EntryMode::DIR => {
            println!("Handling dir {}", entry.path())
        }
        EntryMode::Unknown => continue,
    }
}

Examples

List all entries recursively

This example will list all entries under the dir path/to/dir/

use opendal::EntryMode;
use opendal::Metakey;
use opendal::Operator;
let mut entries = op.list_with("path/to/dir/").recursive(true).await?;
for entry in entries {
    match entry.metadata().mode() {
        EntryMode::FILE => {
            println!("Handling file")
        }
        EntryMode::DIR => {
            println!("Handling dir like start a new list via meta.path()")
        }
        EntryMode::Unknown => continue,
    }
}

List all entries start with prefix

This example will list all entries starts with prefix path/to/prefix

use opendal::EntryMode;
use opendal::Metakey;
use opendal::Operator;
let mut entries = op.list_with("path/to/prefix").recursive(true).await?;
for entry in entries {
    match entry.metadata().mode() {
        EntryMode::FILE => {
            println!("Handling file")
        }
        EntryMode::DIR => {
            println!("Handling dir like start a new list via meta.path()")
        }
        EntryMode::Unknown => continue,
    }
}

pub async fn lister(&self, path: &str) -> Result<Lister>

List entries that starts with given path in parent dir.

This function will create a new [Operator::Lister] to list entries. Users can stop listing via dropping this [Operator::Lister].

Notes

Recursively list

This function only read the children of the given directory. To read all entries recursively, use Operator::lister_with and recursive(true) instead.

Reuse Metadata

The only metadata that is guaranteed to be available is the Mode. For fetching more metadata, please use Operator::lister_with and metakey.

Examples

use futures::TryStreamExt;
use opendal::EntryMode;
use opendal::Metakey;
use opendal::Operator;
let mut ds = op.lister("path/to/dir/").await?;
while let Some(mut de) = ds.try_next().await? {
    match de.metadata().mode() {
        EntryMode::FILE => {
            println!("Handling file")
        }
        EntryMode::DIR => {
            println!("Handling dir like start a new list via meta.path()")
        }
        EntryMode::Unknown => continue,
    }
}

pub fn lister_with( &self, path: &str ) -> FutureLister<impl Future<Output = Result<Lister>>>

List entries that starts with given path in parent dir with options.

This function will create a new [Operator::Lister] to list entries. Users can stop listing via dropping this [Operator::Lister].

Options

start_after

Specify the specified key to start listing from.

This feature can be used to resume a listing from a previous point.

The following example will resume the list operation from the breakpoint.

use opendal::Operator;
let mut lister = op
    .lister_with("path/to/dir/")
    .start_after("breakpoint")
    .await?;

recursive

Specify whether to list recursively or not.

If recursive is set to true, we will list all entries recursively. If not, we’ll only list the entries in the specified dir.

use opendal::Operator;
let mut lister = op.lister_with("path/to/dir/").recursive(true).await?;

metakey

Specify the metadata that required to be fetched in entries.

If metakey is not set, we will fetch only the entry’s mode. Otherwise, we will retrieve the required metadata from storage services. Even if metakey is specified, the metadata may still be None, indicating that the storage service does not supply this information.

Some storage services like s3 could return more metadata like content-length and last-modified. By using metakey, we can fetch those metadata without an extra stat call. Please pick up the metadata you need to reduce the extra stat cost.

This example shows how to list entries with content-length and last-modified metadata:

use futures::TryStreamExt;
use opendal::EntryMode;
use opendal::Metakey;
use opendal::Operator;
let mut lister = op
    .lister_with("dir/")
    // Make sure content-length and last-modified been fetched.
    .metakey(Metakey::ContentLength | Metakey::LastModified)
    .await?;
while let Some(mut entry) = lister.try_next().await? {
    let meta = entry.metadata();
    match meta.mode() {
        EntryMode::FILE => {
            println!(
                "Handling file {} with size {}",
                entry.path(),
                meta.content_length()
            )
        }
        EntryMode::DIR => {
            println!("Handling dir {}", entry.path())
        }
        EntryMode::Unknown => continue,
    }
}

Examples

List all files recursively

use futures::TryStreamExt;
use opendal::EntryMode;
use opendal::Metakey;
use opendal::Operator;
let mut lister = op.lister_with("path/to/dir/").recursive(true).await?;
while let Some(mut entry) = lister.try_next().await? {
    match entry.metadata().mode() {
        EntryMode::FILE => {
            println!("Handling file {}", entry.path())
        }
        EntryMode::DIR => {
            println!("Handling dir {}", entry.path())
        }
        EntryMode::Unknown => continue,
    }
}

List files with required metadata

use futures::TryStreamExt;
use opendal::EntryMode;
use opendal::Metakey;
use opendal::Operator;
let mut ds = op
    .lister_with("path/to/dir/")
    .metakey(Metakey::ContentLength | Metakey::LastModified)
    .await?;
while let Some(mut entry) = ds.try_next().await? {
    let meta = entry.metadata();
    match meta.mode() {
        EntryMode::FILE => {
            println!(
                "Handling file {} with size {}",
                entry.path(),
                meta.content_length()
            )
        }
        EntryMode::DIR => {
            println!("Handling dir {}", entry.path())
        }
        EntryMode::Unknown => continue,
    }
}

impl Operator

Operator presign API.

pub async fn presign_stat( &self, path: &str, expire: Duration ) -> Result<PresignedRequest>

Presign an operation for stat(head).

Example

use anyhow::Result;
use futures::io;
use opendal::Operator;
use std::time::Duration;

async fn test(op: Operator) -> Result<()> {
    let signed_req = op.presign_stat("test",Duration::from_secs(3600)).await?;
    let req = http::Request::builder()
        .method(signed_req.method())
        .uri(signed_req.uri())
        .body(())?;

pub fn presign_stat_with( &self, path: &str, expire: Duration ) -> FuturePresignStat<impl Future<Output = Result<PresignedRequest>>>

Presign an operation for stat(head).

Example

use anyhow::Result;
use futures::io;
use opendal::Operator;
use std::time::Duration;

async fn test(op: Operator) -> Result<()> {
    let signed_req = op.presign_stat_with("test",Duration::from_secs(3600)).override_content_disposition("attachment; filename=\"othertext.txt\"").await?;

pub async fn presign_read( &self, path: &str, expire: Duration ) -> Result<PresignedRequest>

Presign an operation for read.

Notes

Extra Options

presign_read is a wrapper of Self::presign_read_with without any options. To use extra options like override_content_disposition, please use Self::presign_read_with instead.

Example

use anyhow::Result;
use futures::io;
use opendal::Operator;
use std::time::Duration;

async fn test(op: Operator) -> Result<()> {
    let signed_req = op.presign_read("test.txt", Duration::from_secs(3600)).await?;

• signed_req.method(): GET
• signed_req.uri(): https://s3.amazonaws.com/examplebucket/test.txt?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=access_key_id/20130721/us-east-1/s3/aws4_request&X-Amz-Date=20130721T201207Z&X-Amz-Expires=86400&X-Amz-SignedHeaders=host&X-Amz-Signature=<signature-value>
• signed_req.headers(): { "host": "s3.amazonaws.com" }

We can download this file via curl or other tools without credentials:

curl "https://s3.amazonaws.com/examplebucket/test.txt?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=access_key_id/20130721/us-east-1/s3/aws4_request&X-Amz-Date=20130721T201207Z&X-Amz-Expires=86400&X-Amz-SignedHeaders=host&X-Amz-Signature=<signature-value>" -O /tmp/test.txt

pub fn presign_read_with( &self, path: &str, expire: Duration ) -> FuturePresignRead<impl Future<Output = Result<PresignedRequest>>>

Presign an operation for read with extra options.

Options

override_content_disposition

Override the content-disposition header returned by storage services.

use std::time::Duration;

use anyhow::Result;
use opendal::Operator;

async fn test(op: Operator) -> Result<()> {
    let signed_req = op
        .presign_read_with("test.txt", Duration::from_secs(3600))
        .override_content_disposition("attachment; filename=\"othertext.txt\"")
        .await?;
    Ok(())
}

override_cache_control

Override the cache-control header returned by storage services.

use std::time::Duration;

use anyhow::Result;
use opendal::Operator;

async fn test(op: Operator) -> Result<()> {
    let signed_req = op
        .presign_read_with("test.txt", Duration::from_secs(3600))
        .override_cache_control("no-store")
        .await?;
    Ok(())
}

override_content_type

Override the content-type header returned by storage services.

use std::time::Duration;

use anyhow::Result;
use futures::io;
use opendal::Operator;

async fn test(op: Operator) -> Result<()> {
    let signed_req = op
        .presign_read_with("test.txt", Duration::from_secs(3600))
        .override_content_type("text/plain")
        .await?;
    Ok(())
}

pub async fn presign_write( &self, path: &str, expire: Duration ) -> Result<PresignedRequest>

Presign an operation for write.

Notes

Extra Options

presign_write is a wrapper of Self::presign_write_with without any options. To use extra options like content_type, please use Self::presign_write_with instead.

Example

use std::time::Duration;

use anyhow::Result;
use opendal::Operator;

async fn test(op: Operator) -> Result<()> {
    let signed_req = op
        .presign_write("test.txt", Duration::from_secs(3600))
        .await?;
    Ok(())
}

• signed_req.method(): PUT
• signed_req.uri(): https://s3.amazonaws.com/examplebucket/test.txt?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=access_key_id/20130721/us-east-1/s3/aws4_request&X-Amz-Date=20130721T201207Z&X-Amz-Expires=86400&X-Amz-SignedHeaders=host&X-Amz-Signature=<signature-value>
• signed_req.headers(): { "host": "s3.amazonaws.com" }

We can upload file as this file via curl or other tools without credential:

curl -X PUT "https://s3.amazonaws.com/examplebucket/test.txt?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=access_key_id/20130721/us-east-1/s3/aws4_request&X-Amz-Date=20130721T201207Z&X-Amz-Expires=86400&X-Amz-SignedHeaders=host&X-Amz-Signature=<signature-value>" -d "Hello, World!"

pub fn presign_write_with( &self, path: &str, expire: Duration ) -> FuturePresignWrite<impl Future<Output = Result<PresignedRequest>>>

Presign an operation for write with extra options.

Options

content_type

Set the content-type header returned by storage services.

use std::time::Duration;

use anyhow::Result;
use opendal::Operator;

async fn test(op: Operator) -> Result<()> {
    let signed_req = op
        .presign_write_with("test", Duration::from_secs(3600))
        .content_type("text/csv")
        .await?;
    let req = http::Request::builder()
        .method(signed_req.method())
        .uri(signed_req.uri())
        .body(())?;

    Ok(())
}

content_disposition

Set the content-disposition header returned by storage services.

use std::time::Duration;

use anyhow::Result;
use opendal::Operator;

async fn test(op: Operator) -> Result<()> {
    let signed_req = op
        .presign_write_with("test", Duration::from_secs(3600))
        .content_disposition("attachment; filename=\"cool.html\"")
        .await?;
    let req = http::Request::builder()
        .method(signed_req.method())
        .uri(signed_req.uri())
        .body(())?;

    Ok(())
}

cache_control

Set the cache-control header returned by storage services.

use std::time::Duration;

use anyhow::Result;
use opendal::Operator;

async fn test(op: Operator) -> Result<()> {
    let signed_req = op
        .presign_write_with("test", Duration::from_secs(3600))
        .cache_control("no-store")
        .await?;
    let req = http::Request::builder()
        .method(signed_req.method())
        .uri(signed_req.uri())
        .body(())?;

    Ok(())
}

impl Operator

Operator build API

Operator should be built via OperatorBuilder. We recommend to use Operator::new to get started:

use opendal::services::Fs;
use opendal::Operator;
async fn test() -> Result<()> {
    // Create fs backend builder.
    let mut builder = Fs::default();
    // Set the root for fs, all operations will happen under this root.
    //
    // NOTE: the root must be absolute path.
    builder.root("/tmp");

    // Build an `Operator` to start operating the storage.
    let op: Operator = Operator::new(builder)?.finish();

    Ok(())
}

pub fn new<B: Builder>(ab: B) -> Result<OperatorBuilder<impl Accessor>>

Create a new operator with input builder.

OpenDAL will call builder.build() internally, so we don’t need to import opendal::Builder trait.

Examples

Read more backend init examples in examples.

use opendal::services::Fs;
use opendal::Operator;
async fn test() -> Result<()> {
    // Create fs backend builder.
    let mut builder = Fs::default();
    // Set the root for fs, all operations will happen under this root.
    //
    // NOTE: the root must be absolute path.
    builder.root("/tmp");

    // Build an `Operator` to start operating the storage.
    let op: Operator = Operator::new(builder)?.finish();

    Ok(())
}

pub fn from_map<B: Builder>( map: HashMap<String, String> ) -> Result<OperatorBuilder<impl Accessor>>

Create a new operator from given map.

Notes

from_map is using static dispatch layers which is zero cost. via_map is using dynamic dispatch layers which has a bit runtime overhead with an extra vtable lookup and unable to inline. But from_map requires generic type parameter which is not always easy to be used.

Examples

use std::collections::HashMap;

use opendal::services::Fs;
use opendal::Operator;
async fn test() -> Result<()> {
    let map = HashMap::from([
        // Set the root for fs, all operations will happen under this root.
        //
        // NOTE: the root must be absolute path.
        ("root".to_string(), "/tmp".to_string()),
    ]);

    // Build an `Operator` to start operating the storage.
    let op: Operator = Operator::from_map::<Fs>(map)?.finish();

    Ok(())
}

pub fn via_map(scheme: Scheme, map: HashMap<String, String>) -> Result<Operator>

Create a new operator from given scheme and map.

Notes

from_map is using static dispatch layers which is zero cost. via_map is using dynamic dispatch layers which has a bit runtime overhead with an extra vtable lookup and unable to inline. But from_map requires generic type parameter which is not always easy to be used.

Examples

use std::collections::HashMap;

use opendal::Operator;
use opendal::Scheme;
async fn test() -> Result<()> {
    let map = HashMap::from([
        // Set the root for fs, all operations will happen under this root.
        //
        // NOTE: the root must be absolute path.
        ("root".to_string(), "/tmp".to_string()),
    ]);

    // Build an `Operator` to start operating the storage.
    let op: Operator = Operator::via_map(Scheme::Fs, map)?;

    Ok(())
}

pub fn layer<L: Layer<FusedAccessor>>(self, layer: L) -> Self

Create a new layer with dynamic dispatch.

Notes

OperatorBuilder::layer() is using static dispatch which is zero cost. Operator::layer() is using dynamic dispatch which has a bit runtime overhead with an extra vtable lookup and unable to inline.

It’s always recommended to use OperatorBuilder::layer() instead.

Examples

use opendal::layers::LoggingLayer;
use opendal::services::Fs;
use opendal::Operator;

let op = Operator::new(Fs::default())?.finish();
let op = op.layer(LoggingLayer::default());
// All operations will go through the new_layer
let _ = op.read("test_file").await?;

Trait Implementations

impl Clone for Operator

fn clone(&self) -> Operator

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for Operator

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl From<BlockingOperator> for Operator

fn from(v: BlockingOperator) -> Self

Converts to this type from the input type.