import ChangeLog from ‘../changelog/connector-file-sftp.md’;

SftpFile

Sftp file source connector

Support Those Engines

Spark
Flink
SeaTunnel Zeta

Key Features

[x] batch
[ ] stream
[x] multimodal
Use binary file format to read and write files in any format, such as videos, pictures, etc. In short, any files can be synchronized to the target place.
[ ] exactly-once
[x] column projection
[x] parallelism
[ ] support user-defined split
[x] file format type
- [x] text
- [x] csv
- [x] json
- [x] excel
- [x] xml
- [x] binary
- [x] markdown

Description

Read data from sftp file server.

Supported DataSource Info

In order to use the SftpFile connector, the following dependencies are required. They can be downloaded via install-plugin.sh or from the Maven central repository.

Datasource	Supported Versions	Dependency
SftpFile	universal	Download

:::tip

If you use spark/flink, In order to use this connector, You must ensure your spark/flink cluster already integrated hadoop. The tested hadoop version is 2.x.

If you use SeaTunnel Engine, It automatically integrated the hadoop jar when you download and install SeaTunnel Engine. You can check the jar package under ${SEATUNNEL_HOME}/lib to confirm this.

We made some trade-offs in order to support more file types, so we used the HDFS protocol for internal access to Sftp and this connector need some hadoop dependencies. It only supports hadoop version 2.9.X+.

:::

Data Type Mapping

The File does not have a specific type list, and we can indicate which SeaTunnel data type the corresponding data needs to be converted to by specifying the Schema in the config.

SeaTunnel Data type
STRING
SHORT
INT
BIGINT
BOOLEAN
DOUBLE
DECIMAL
FLOAT
DATE
TIME
TIMESTAMP
BYTES
ARRAY
MAP

Source Options

Name	Type	Required	default value	Description
host	String	Yes	-	The target sftp host is required
port	Int	Yes	-	The target sftp port is required
user	String	Yes	-	The target sftp username is required
password	String	No	-	The target sftp password. Required when `keyfile` is not set.
keyfile	String	No	-	The private key file path used for SFTP public key authentication.
path	String	Yes	-	The source file path.
file_format_type	String	Yes	-	Please check #file_format_type below
file_filter_pattern	String	No	-	Filter pattern, which used for filtering files.
filename_extension	string	no	-	Filter filename extension, which used for filtering files with specific extension. Example: `csv` `.txt` `json` `.xml`.
delimiter/field_delimiter	String	No	\001 for text and ‘,’ for csv	delimiter parameter will deprecate after version 2.3.5, please use field_delimiter instead. Field delimiter, used to tell connector how to slice and dice fields when reading text files. Default `\001`, the same as hive's default delimiter
row_delimiter	string	no	\n	Row delimiter, used to tell connector how to slice and dice rows when reading text files. Default `\n`
parse_partition_from_path	Boolean	No	true	Control whether parse the partition keys and values from file path For example if you read a file from path `oss://hadoop-cluster/tmp/seatunnel/parquet/name=tyrantlucifer/age=26` Every record data from file will be added these two fields: name age tyrantlucifer 26 Tips: Do not define partition fields in schema option
date_format	String	No	yyyy-MM-dd	Date type format, used to tell connector how to convert string to date, supported as the following formats: `yyyy-MM-dd` `yyyy.MM.dd` `yyyy/MM/dd` default `yyyy-MM-dd`
datetime_format	String	No	yyyy-MM-dd HH:mm:ss	Datetime type format, used to tell connector how to convert string to datetime, supported as the following formats: `yyyy-MM-dd HH:mm:ss` `yyyy.MM.dd HH:mm:ss` `yyyy/MM/dd HH:mm:ss` `yyyyMMddHHmmss` default `yyyy-MM-dd HH:mm:ss`
time_format	String	No	HH:mm:ss	Time type format, used to tell connector how to convert string to time, supported as the following formats: `HH:mm:ss` `HH:mm:ss.SSS` default `HH:mm:ss`
skip_header_row_number	Long	No	0	Skip the first few lines, but only for the txt and csv. For example, set like following: `skip_header_row_number = 2` then SeaTunnel will skip the first 2 lines from source files
read_columns	list	no	-	The read column list of the data source, user can use it to implement field projection.
sheet_name	String	No	-	Reader the sheet of the workbook,Only used when file_format is excel.
xml_row_tag	string	no	-	Specifies the tag name of the data rows within the XML file, only used when file_format is xml.
xml_use_attr_format	boolean	no	-	Specifies whether to process data using the tag attribute format, only used when file_format is xml.
csv_use_header_line	boolean	no	false	Whether to use the header line to parse the file, only used when the file_format is `csv` and the file contains the header line that match RFC 4180
schema	Config	No	-	Please check #schema below
compress_codec	String	No	None	The compress codec of files and the details that supported as the following shown: - txt: `lzo` `None` - json: `lzo` `None` - csv: `lzo` `None` - orc: `lzo` `snappy` `lz4` `zlib` `None` - parquet: `lzo` `snappy` `lz4` `gzip` `brotli` `zstd` `None` Tips: excel type does Not support any compression format
archive_compress_codec	string	no	none
encoding	string	no	UTF-8
null_format	string	no	-	Only used when file_format_type is text. null_format to define which strings can be represented as null. e.g: `\N`
binary_chunk_size	int	no	1024	Only used when file_format_type is binary. The chunk size (in bytes) for reading binary files. Default is 1024 bytes. Larger values may improve performance for large files but use more memory.
binary_complete_file_mode	boolean	no	false	Only used when file_format_type is binary. Whether to read the complete file as a single chunk instead of splitting into chunks. When enabled, the entire file content will be read into memory at once. Default is false.
discovery_mode	string	no	once	File discovery mode. Supported values: `once` (default), `continuous`. When `continuous`, the source keeps scanning the path and processes new/changed files at runtime (unbounded). In the current implementation, `continuous` requires `sync_mode=update` (binary only).
scan_interval	string	no	10S	Only used when `discovery_mode=continuous`. Scan interval for periodic discovery, recommended shorthand format `10S`, `30S`; ISO-8601 format `PT10S`, `PT30S` is also supported.
start_mode	string	no	earliest	Only used when `discovery_mode=continuous`. Supported values: `earliest` (default), `latest`.
sync_mode	string	no	full	File sync mode. Supported values: `full`, `update`. When `update`, the source compares files between source/target and only reads new/changed files (currently only supports `file_format_type=binary`).
target_path	string	no	-	Only used when `sync_mode=update`. Target base path used for comparison (it should usually be the same as sink `path`).
target_hadoop_conf	map	no	-	Only used when `sync_mode=update`. Extra Hadoop configuration for target filesystem. You can set `fs.defaultFS` in this map to override target defaultFS.
update_strategy	string	no	distcp	Only used when `sync_mode=update`. Supported values: `distcp` (default), `strict`.
compare_mode	string	no	len_mtime	Only used when `sync_mode=update`. Supported values: `len_mtime` (default), `checksum` (only valid when `update_strategy=strict`).
common-options		No	-	Source plugin common parameters, please refer to Source Common Options for details.
file_filter_modified_start	string	no	-	File modification time filter. The connector will filter some files base on the last modification start time (include start time). The default data format is `yyyy-MM-dd HH:mm:ss`.
file_filter_modified_end	string	no	-	File modification time filter. The connector will filter some files base on the last modification end time (not include end time). The default data format is `yyyy-MM-dd HH:mm:ss`.
quote_char	string	no	"	A single character that encloses CSV fields, allowing fields with commas, line breaks, or quotes to be read correctly.
escape_char	string	no	-	A single character that allows the quote or other special characters to appear inside a CSV field without ending the field.
metalake_type	string	no	gravitino	The type of metalake service, currently supports `gravitino`.
recursive_file_scan	boolean	no	true	Whether to scan subdirectories recursively. If `false`, subdirectories will be ignored.
sort_files_by_modification_time	boolean	no	false	Sort files by modification time in descending order. Enable this when reading evolving schemas to ensure schema inference uses the latest file.

file_filter_pattern [string]

Filter pattern, which used for filtering files. If you only want to filter based on file names, simply write the regular file names; If you want to filter based on the file directory at the same time, the expression needs to start with path.

The pattern follows standard regular expressions. For details, please refer to https://en.wikipedia.org/wiki/Regular_expression. There are some examples.

If the path is /data/seatunnel, and the file structure example is:

/data/seatunnel/20241001/report.txt
/data/seatunnel/20241007/abch202410.csv
/data/seatunnel/20241002/abcg202410.csv
/data/seatunnel/20241005/old_data.csv
/data/seatunnel/20241012/logo.png

Matching Rules Example:

Example 1: Match all .txt files，Regular Expression:

.*.txt

The result of this example matching is:

/data/seatunnel/20241001/report.txt

Example 2: Match all file starting with abc，Regular Expression:

abc.*

The result of this example matching is:

/data/seatunnel/20241007/abch202410.csv
/data/seatunnel/20241002/abcg202410.csv

Example 3: Match all files starting with abc in folder 20241007，And the fourth character is either h or g, the Regular Expression:

/data/seatunnel/20241007/abc[h,g].*

The result of this example matching is:

/data/seatunnel/20241007/abch202410.csv

Example 4: Match third level folders starting with 202410 and files ending with .csv, the Regular Expression:

/data/seatunnel/202410\d*/.*.csv

The result of this example matching is:

/data/seatunnel/20241007/abch202410.csv
/data/seatunnel/20241002/abcg202410.csv
/data/seatunnel/20241005/old_data.csv

file_format_type [string]

File type, supported as the following file types: text csv parquet orc json excel xml binary markdown If you assign file type to json, you should also assign schema option to tell connector how to parse data to the row you want. For example: upstream data is the following:

{"code":  200, "data":  "get success", "success":  true}

You can also save multiple pieces of data in one file and split them by newline:

{"code":  200, "data":  "get success", "success":  true}
{"code":  300, "data":  "get failed", "success":  false}

you should assign schema as the following:

schema {
    fields {
        code = int
        data = string
        success = boolean
    }
}

connector will generate data as the following:

code	data	success
200	get success	true
If you assign file type to `parquet` `orc`, schema option not required, connector can find the schema of upstream data automatically.
If you assign file type to `text` `csv`, you can choose to specify the schema information or not.
For example, upstream data is the following:

tyrantlucifer#26#male

If you do not assign data schema connector will treat the upstream data as the following:

content
tyrantlucifer#26#male
If you assign data schema, you should also assign the option `field_delimiter` too except CSV file type
you should assign schema and delimiter as the following:

field_delimiter = "#"
schema {
    fields {
        name = string
        age = int
        gender = string 
    }
}

connector will generate data as the following:

name	age	gender
tyrantlucifer	26	male

If you assign file type to binary, SeaTunnel can synchronize files in any format, such as compressed packages, pictures, etc. In short, any files can be synchronized to the target place. Under this requirement, you need to ensure that the source and sink use binary format for file synchronization at the same time.

If you assign file type to markdown, SeaTunnel can parse markdown files and extract structured data. The markdown parser extracts various elements including headings, paragraphs, lists, code blocks, tables, and more. Each element is converted to a row with the following schema:

element_id: Unique identifier for the element
element_type: Type of the element (Heading, Paragraph, ListItem, etc.)
heading_level: Level of heading (1-6, null for non-heading elements)
text: Text content of the element
page_number: Page number (default: 1)
position_index: Position index within the document
parent_id: ID of the parent element
child_ids: Comma-separated list of child element IDs

When markdown_rag_metadata_enabled is set to true, SeaTunnel appends the following RAG metadata fields after child_ids:

source_uri: Source file path or URI
document_id: Stable document identifier derived from source_uri
chunk_id: Stable chunk identifier derived from document identity, chunk order, and content hash
chunk_index: One-based chunk order in the parsed document
content_hash: SHA-256 hash of the emitted text value

The option defaults to false, so the original Markdown schema is unchanged unless you enable it.

Note: Markdown format only supports reading, not writing.

compress_codec [string]

The compress codec of files and the details that supported as the following shown:

txt: lzo none
json: lzo none
csv: lzo none
orc/parquet:
automatically recognizes the compression type, no additional settings required.

archive_compress_codec [string]

The compress codec of archive files and the details that supported as the following shown:

archive_compress_codec	file_format	archive_compress_suffix
ZIP	txt,json,excel,xml	.zip
TAR	txt,json,excel,xml	.tar
TAR_GZ	txt,json,excel,xml	.tar.gz
GZ	txt,json,excel,xml	.gz
NONE	all	.*

Note: gz compressed excel file needs to compress the original file or specify the file suffix, such as e2e.xls ->e2e_test.xls.gz

encoding [string]

Only used when file_format_type is json,text,csv,xml. The encoding of the file to read. This param will be parsed by Charset.forName(encoding).

binary_chunk_size [int]

Only used when file_format_type is binary.

The chunk size (in bytes) for reading binary files. Default is 1024 bytes. Larger values may improve performance for large files but use more memory.

binary_complete_file_mode [boolean]

Only used when file_format_type is binary.

Whether to read the complete file as a single chunk instead of splitting into chunks. When enabled, the entire file content will be read into memory at once. Default is false.

discovery_mode [string]

File discovery mode. Supported values: once (default), continuous.

once: enumerate current files once and finish (bounded).
continuous: keep scanning the path and processing new/changed files at runtime (unbounded).

In the current implementation, discovery_mode=continuous requires sync_mode=update (binary only) to avoid repeated transfers.

scan_interval [string]

Only used when discovery_mode=continuous. Scan interval for periodic discovery; value must be greater than 0. Recommended shorthand format 10S, 30S (case-insensitive, e.g. 10s); ISO-8601 format PT10S, PT30S is also supported. Default is 10S.

start_mode [string]

Only used when discovery_mode=continuous. Supported values: earliest (default), latest.

earliest: read existing files on startup.
latest: only process files modified after the job starts.

sync_mode [string]

File sync mode. Supported values: full (default), update. When update, the source compares files between source/target and only reads new/changed files (currently only supports file_format_type=binary).

Performance considerations

Update mode triggers an extra getFileStatus call on the target for each source file.
For remote file systems (FTP/SFTP), this adds per-file network overhead. It is not recommended for massive small-file scenarios.

Requirements / limitations

target_path should typically align with sink path (same filesystem and same relative path layout).
When update_strategy=distcp, correctness depends on source/target clock synchronization.
When compare_mode=checksum, filesystem checksum support is required. If checksum is unavailable, SeaTunnel falls back to content comparison (more expensive) and logs a warning.

Example:

sync_mode = "update"
file_format_type = "binary"
target_path = "/path/to/your/sink/path"
update_strategy = "distcp"
compare_mode = "len_mtime"

target_path [string]

Only used when sync_mode=update. Target base path used for comparison (it should usually be the same as sink path).

target_hadoop_conf [map]

Only used when sync_mode=update. Extra Hadoop configuration for target filesystem. You can set fs.defaultFS in this map to override target defaultFS.

update_strategy [string]

Only used when sync_mode=update. Supported values: distcp (default), strict.

compare_mode [string]

Only used when sync_mode=update. Supported values: len_mtime (default), checksum (only valid when update_strategy=strict).

quote_char [string]

A single character that encloses CSV fields, allowing fields with commas, line breaks, or quotes to be read correctly.

escape_char [string]

A single character that allows the quote or other special characters to appear inside a CSV field without ending the field.

recursive_file_scan [boolean]

Whether to scan subdirectories recursively. If false, subdirectories will be ignored.

sort_files_by_modification_time [boolean]

Whether to sort files by modification time in descending order. Default is false.

When enabled, files will be sorted by their modification time (newest first). This is useful when:

Reading files with evolving schemas and you want schema inference to use the latest file
You need to process files in chronological order

schema [config]

fields [Config]

The schema of upstream data. For more details, please refer to Schema Feature.

metadata_table_id [string]

The table identifier in the metadata service to fetch table schema. For Gravitino, the format should be {catalog}.{database}.{table}, such as mysql-catalog.test_db.users.

When specified, the connector will fetch table schema from the external metadata service instead of using manual columns definition.

When using Gravitino as the metadata source, the column types from Gravitino will be automatically converted to SeaTunnel data types. For detailed type mapping information, please refer to Gravitino Type Mapping.

For more information, please refer to Metadata SPI.

How to Create a Sftp Data Synchronization Jobs

The following example demonstrates how to create a data synchronization job that reads data from sftp and prints it on the local client:

# Set the basic configuration of the task to be performed
env {
  parallelism = 1
  job.mode = "BATCH"
}

# Create a source to connect to sftp
source {
  SftpFile {
    host = "sftp"
    port = 22
    user = seatunnel
    password = pass
    path = "tmp/seatunnel/read/json"
    file_format_type = "json"
    plugin_output = "sftp"
    schema = {
      fields {
        c_map = "map<string, string>"
        c_array = "array<int>"
        c_string = string
        c_boolean = boolean
        c_tinyint = tinyint
        c_smallint = smallint
        c_int = int
        c_bigint = bigint
        c_float = float
        c_double = double
        c_bytes = bytes
        c_date = date
        c_decimal = "decimal(38, 18)"
        c_timestamp = timestamp
        c_row = {
          C_MAP = "map<string, string>"
          C_ARRAY = "array<int>"
          C_STRING = string
          C_BOOLEAN = boolean
          C_TINYINT = tinyint
          C_SMALLINT = smallint
          C_INT = int
          C_BIGINT = bigint
          C_FLOAT = float
          C_DOUBLE = double
          C_BYTES = bytes
          C_DATE = date
          C_DECIMAL = "decimal(38, 18)"
          C_TIMESTAMP = timestamp
        }
      }
    }
  }
}

# Console printing of the read sftp data
sink {
  Console {
    parallelism = 1
  }
}

Multiple Table


SftpFile {
  tables_configs = [
    {
      schema {
        table = "student"
        fields {
          name = string
          age = int
        }
      }
      path = "/tmp/seatunnel/sink/text"
      host = "192.168.31.48"
      port = 21
      user = tyrantlucifer
      password = tianchao
      file_format_type = "parquet"
    },
    {
      schema {
        table = "teacher"
        fields {
          name = string
          age = int
        }
      }
      path = "/tmp/seatunnel/sink/text"
      host = "192.168.31.48"
      port = 21
      user = tyrantlucifer
      password = tianchao
      file_format_type = "parquet"
    }
  ]
}

Filter File

env {
  parallelism = 1
  job.mode = "BATCH"
}

source {
  SftpFile {
    host = "sftp"
    port = 22
    user = seatunnel
    password = pass
    path = "tmp/seatunnel/read/json"
    file_format_type = "json"
    plugin_output = "sftp"
    // file example abcD2024.csv
    file_filter_pattern = "abc[DX]*.*"
  }
}

sink {
  Console {
  }
}

Incremental Sync (sync_mode=update, binary)

sync_mode=update compares files between source and target_path, then only reads new/changed files. In most cases, target_path should be aligned with sink path (same filesystem and same relative paths).

env {
  parallelism = 1
  job.mode = "BATCH"
}

source {
  SftpFile {
    host = "sftp"
    port = 22
    user = seatunnel
    password = pass

    path = "tmp/seatunnel/update/src"
    file_format_type = "binary"

    sync_mode = "update"
    target_path = "tmp/seatunnel/update/dst"
    update_strategy = "distcp"
    compare_mode = "len_mtime"
  }
}

sink {
  SftpFile {
    host = "sftp"
    port = 22
    user = seatunnel
    password = pass

    path = "tmp/seatunnel/update/dst"
    tmp_path = "tmp/seatunnel/update/tmp"
    file_format_type = "binary"
  }
}

Continuous Discovery (discovery_mode=continuous)

discovery_mode=continuous keeps the job running and periodically scans the path for new/changed files (long-running job, recommended to run with job.mode="STREAMING").

Note: discovery_mode=continuous currently requires sync_mode="update" (binary-only) to avoid repeated transfers without keeping an unbounded “seen” state. target_path should align with the sink path on the same filesystem.

env {
  parallelism = 1
  job.mode = "STREAMING"
}

source {
  SftpFile {
    host = "sftp"
    port = 22
    user = seatunnel
    password = pass

    path = "tmp/seatunnel/watch/src"
    file_format_type = "binary"

    discovery_mode = "continuous"
    scan_interval = "10S"
    start_mode = "latest"

    sync_mode = "update"
    target_path = "tmp/seatunnel/watch/dst"
    update_strategy = "distcp"
    compare_mode = "len_mtime"
  }
}

sink {
  SftpFile {
    host = "sftp"
    port = 22
    user = seatunnel
    password = pass

    path = "tmp/seatunnel/watch/dst"
    tmp_path = "tmp/seatunnel/watch/tmp"
    file_format_type = "binary"
  }
}

SftpFile

Support Those Engines

Key Features

Description

Supported DataSource Info

Data Type Mapping

Source Options

file_filter_pattern [string]

file_format_type [string]

compress_codec [string]

archive_compress_codec [string]

encoding [string]

binary_chunk_size [int]

binary_complete_file_mode [boolean]

discovery_mode [string]

scan_interval [string]

start_mode [string]

sync_mode [string]

target_path [string]

target_hadoop_conf [map]

update_strategy [string]

compare_mode [string]

quote_char [string]

escape_char [string]

recursive_file_scan [boolean]

sort_files_by_modification_time [boolean]

schema [config]

fields [Config]

metadata_table_id [string]

How to Create a Sftp Data Synchronization Jobs

Multiple Table

Filter File

Incremental Sync (sync_mode=update, binary)

Continuous Discovery (discovery_mode=continuous)

Changelog