File Get¶

Description¶

File Get is a source node that downloads a file from an SFTP location and makes the file available to downstream nodes.

Use File Get when a pipeline needs to retrieve a source file before loading, transforming, validating, or delivering the data. By default, File Get passes the downloaded file forward as a stream handoff. It can also publish a durable artifact copy when the pipeline needs to preserve the downloaded file separately.

Field reference¶

Field	Description
Protocol	Required. File transfer protocol to use. File Get currently supports `sftp` only.
Account	Required. Account UUID used for SFTP authentication. The selected account must contain a username and either password-based or SSH key-based credentials.
Host	Required. SFTP host name or IP address. Enter the host only. Do not include `sftp://`, a path, slashes, backslashes, or whitespace.
Port	Optional. SFTP port number. Defaults to `22`. The value must be an integer between `1` and `65535`.
Remote path	Required. Remote file path to download, such as `/outgoing/data.csv`. Use forward slashes. Backslashes and null bytes are rejected. This field supports expressions.
Max bytes	Optional. Maximum number of bytes File Get may download. The configured value can lower, but not raise, the server-level File Get limit.
Output filename	Optional. Filename used for the stream packet and optional artifact copy. When blank, SchemAlign uses the remote file name or a safe fallback name. This field supports expressions.
Format	Optional. How SchemAlign should classify the downloaded file for metadata and downstream consumers. Supported values are `auto`, `csv`, `json`, `xml`, and `txt`. Defaults to `auto`.
Also publish artifact copy	Optional. When enabled, SchemAlign writes a durable artifact copy in addition to the normal stream handoff. Defaults to off.
Keep artifact history	Optional. When artifact publishing is enabled, controls whether prior artifact versions are retained. Defaults to on.
Allow global account	Optional. Allows a global account to be selected when the project permits global accounts. Defaults to disabled.

When to use File Get¶

Use this node when the pipeline should:

download a file from an SFTP server
pass the downloaded file directly to a downstream SQL Loader, SQL Transform, or destination node
preserve the original downloaded file as an optional artifact copy
enforce a maximum download size
use runtime values in the remote path or output filename

File Get is a source node. It normally appears at the beginning of a pipeline branch or after a node that supplies a single-row parameter payload.

Common pipeline patterns¶

Pattern	Use case
`File Get -> SQL Loader -> SQL Transform`	Download a file, load it into SQL-accessible working data, then transform the result.
`File Get -> SQL Transform`	Download a file and pass it to a transform path that can consume the stream handoff.
`File Get -> File Put`	Copy or relay a file from one configured location to another.
`File Get -> SQL Transform -> HTTP Response`	Download and transform a file as part of a request-driven workflow.

Expressions and upstream placeholders¶

File Get supports expressions in these fields:

Field	Expression result type	Example
Remote path	String	`/exports/{{yyyymmdd}}/input.csv`
Output filename	String	`input_{{run_id}}.csv`

File Get also supports single-brace upstream placeholders when it is connected to exactly one upstream node that produces a single-row parameter payload. For example, a prior node can publish a value named file_name, and File Get can use a remote path such as /exports/{file_name}.

Supported upstream parameter payload formats include text, CSV, TSV, JSON, and NDJSON when the payload resolves to one record or one scalar value. Multiple upstream nodes, multiple rows, or unsupported formats produce validation errors.

Authentication¶

File Get requires an SFTP account. The selected account must be available to the current project and must contain valid credentials.

Supported credential patterns are:

Account type	Required credential details
Basic authentication	Username and password.
SSH key	Username and private key. A key passphrase may also be supplied when required by the key.

Do not place usernames, passwords, private keys, or key passphrases in Host, Remote path, Output filename, or expression values. Store credentials in an Account instead.

Output and artifact behavior¶

File Get always attempts to publish a stream handoff for downstream connected nodes.

Output	Description
Stream packet	The normal downstream handoff containing downloaded bytes, inferred or selected format, MIME type, filename, and source metadata.
Artifact copy	Optional durable copy created only when Also publish artifact copy is enabled.
Run output metadata	Includes status, mode, format, filename, MIME type, max byte limit, account reference, and remote path.

When artifact publishing is enabled, File Get writes the downloaded file into the run directory first, then publishes it through the artifact system. When artifact publishing is disabled, downstream connected nodes should consume the stream handoff directly.

Format handling¶

When Format is set to auto, SchemAlign infers the format from the remote path extension:

Remote path ending	Inferred format
`.json`	`json`
`.xml`	`xml`
`.csv`	`csv`
`.tsv`	`csv`
`.txt`	`csv`
Other or unknown	`csv`

Set Format explicitly when the extension does not match how downstream nodes should treat the file.

Validation and common errors¶

Message or condition	What it means	How to fix it
`file_get requires config.account_id`	No account was selected.	Select an SFTP account that is available to the project.
`file_get requires config.host`	Host is blank.	Enter the SFTP host name or IP address.
`file_get config.host must not include a URL scheme`	Host includes text such as `sftp://`.	Enter only the host name or IP address.
`file_get config.host must not include '/' or '\\'`	Host includes path characters.	Move path content to Remote path.
`file_get config.host must not contain whitespace`	Host contains spaces or line breaks.	Remove whitespace from the host value.
`file_get config.port must be an integer`	Port is not numeric.	Enter a valid port number, usually `22`.
`file_get config.port must be between 1 and 65535`	Port is outside the valid TCP port range.	Enter a port between `1` and `65535`.
`file_get requires config.remote_path`	Remote path is blank.	Enter the remote file path to download.
`file_get config.remote_path must not contain backslashes`	Remote path uses Windows-style separators.	Use forward slashes in the SFTP path.
`file_get config.remote_path is too long`	Remote path exceeds the configured safety limit.	Shorten the remote path.
`file_get currently supports protocol=sftp only`	A protocol other than SFTP was configured.	Use `sftp`.
`Selected account has no username configured`	The account is missing a username.	Update the account credentials.
`Selected account is basic_auth but has no password`	The selected password-based account is incomplete.	Add the password to the account credentials.
`Selected account is ssh_key but has no private_key`	The selected SSH key account is incomplete.	Add the private key to the account credentials.
`Unsupported SSH private key format`	The SSH key could not be parsed.	Use a supported private key format.
`Remote file too large` or `Downloaded file too large`	The file exceeds the effective max byte limit.	Increase the allowed limit if appropriate, or download a smaller file.
`Connector parameter binding supports exactly one upstream node`	The node has more than one upstream dependency for parameter binding.	Connect only one upstream parameter source or remove upstream placeholders.
`Upstream parameter CSV must contain exactly 1 data row`	The upstream parameter payload contains more than one CSV row.	Make the upstream payload resolve to one row.
`Upstream parameter JSON array must contain exactly 1 item`	The upstream JSON payload contains multiple array items.	Make the upstream payload resolve to one item.
`Unsupported upstream parameter payload format`	The upstream payload format cannot be used for placeholder binding.	Use text, CSV, TSV, JSON, or NDJSON for parameter binding.

Security notes¶

Use Accounts for SFTP credentials; do not paste secrets into node fields.
Keep Allow global account disabled unless the project intentionally permits a global account.
Use Max bytes to reduce the chance of accidental large downloads.
Remote path and output filename expressions are runtime configuration and should be reviewed before production use.
File Get enforces project account scoping at runtime before using selected credentials.