DML Commands - Data Loading
Uploads (i.e. stages) data files from a local directory/folder on a client machine to one of the following Snowflake stages:
Named internal stage.
Internal stage for a specified table.
Internal stage for the current user.
Once files are staged, the data in the files can be loaded into a table using the COPY INTO <table> command.
PUT does not support uploading files to external stages. To upload files to external stages, use the utilities provided by the cloud service.
The following Snowflake clients do not support PUT:
The ODBC driver supports PUT with Snowflake accounts hosted on the following platforms:
Amazon Web Services (using ODBC Driver Version 2.17.5 and higher).
Google Cloud Platform (using ODBC Driver Version 2.21.5 and higher).
Microsoft Azure (using ODBC Driver Version 2.20.2 and higher).
PUT file://<path_to_file>/<filename> internalStage [ PARALLEL = <integer> ] [ AUTO_COMPRESS = TRUE | FALSE ] [ SOURCE_COMPRESSION = AUTO_DETECT | GZIP | BZ2 | BROTLI | ZSTD | DEFLATE | RAW_DEFLATE | NONE ] [ OVERWRITE = TRUE | FALSE ]
internalStage ::= @[<namespace>.]<int_stage_name>[/<path>] | @[<namespace>.]%<table_name>[/<path>] | @~[/<path>]
Specifies the URI for the data file(s) on the client machine, where:
path_to_fileis the local directory path to the file(s) to upload. If the files are located in the root directory (or sub-directory) on the client machine:
You must include the initial forward slash in the path (e.g.
You must include the drive and backslash in the path (e.g.
filenameis the name of the file(s) to upload. Wildcard characters (
?) are supported to enable uploading multiple files in a directory.
The URI can be enclosed in single quotes, which allows special characters, including spaces, in directory and file names; however, the drive and path separator is a forward slash (
/) for all supported operating systems (e.g.
'file://C:/temp/load data'for a path in Windows containing a directory named
Specifies the location in Snowflake where to upload the files:
Files are uploaded to the specified named internal stage.
Files are uploaded to the stage for the specified table.
Files are uploaded to the stage for the current user.
namespaceis the database and/or schema in which the named internal stage or table resides. It is optional if a database and schema are currently in use within the session; otherwise, it is required.
pathis an optional case-sensitive path for files in the cloud storage location (i.e. files have names that begin with a common string) that limits access to a set of files. Paths are alternatively called prefixes or folders by different cloud storage services.
If the stage name or path includes spaces or special characters, it must be enclosed in single quotes (e.g.
'@"my stage"'for a stage named
PARALLEL = integer
Specifies the number of threads to use for uploading files. The upload process separate batches of data files by size:
Small files (< 64 MB compressed or uncompressed) are staged in parallel as individual files.
Larger files are automatically split into chunks, staged concurrently, and reassembled in the target stage. A single thread can upload multiple chunks.
Increasing the number of threads can improve performance when uploading large files.
Supported values: Any integer value from
1(no parallelism) to
99(use 99 threads for uploading files).
A 16 MB (rather than 64 MB) limit applies to older versions of Snowflake drivers, including:
JDBC Driver versions prior to 3.12.1.
ODBC Driver versions prior to 2.20.5.
Python Connector versions prior to 2.2.0.
AUTO_COMPRESS = TRUE | FALSE
Specifies whether Snowflake uses gzip to compress files during upload:
TRUE: Files are compressed (if they are not already compressed).
FALSE: Files are not compressed (i.e. files are uploaded as-is).
This option does not support other compression types. To use a different compression type, compress the file separately before executing the PUT command. Then, identify the compression type using the
Ensure your local folder has sufficient space for Snowflake to compress the data files before staging them. If necessary, set the
TMPenvironment variable in your operating system to point to a local folder that contains additional free space.
SOURCE_COMPRESSION = AUTO_DETECT | GZIP | BZ2 | BROTLI | ZSTD | DEFLATE | RAW_DEFLATE | NONE
Specifies the method of compression used on already-compressed files that are being staged:
Compression algorithm detected automatically, except for Brotli-compressed files, which cannot currently be detected automatically. If loading Brotli-compressed files, explicitly use
Must be used if loading Brotli-compressed files.
Zstandard v0.8 (and higher) supported.
Deflate-compressed files (with zlib header, RFC1950).
Raw Deflate-compressed files (without header, RFC1951).
Data files to load have not been compressed.
Snowflake uses this option to detect how the data files were compressed so that they can be uncompressed and the data extracted for loading; it does not use this option to compress the files.
Uploading files that were compressed with other utilities (e.g. lzip, lzma, lzop, and xz) is not currently supported.
OVERWRITE = TRUE | FALSE
Specifies whether Snowflake overwrites an existing file with the same name during upload:
TRUE: An existing file with the same name is overwritten.
FALSE: An existing file with the same name is not overwritten.
Note that a LIST operation on the stage is performed in the background, which can affect the performance of the PUT operation.
If attempts to PUT a file fail because a file with the same name exists in the target stage, the following options are available:
Load the data from the existing file into one or more tables, and remove the file from the stage. Then PUT a file with new or updated data to the stage.
Rename the local file, and then attempt the PUT operation again.
OVERWRITE = TRUEin the PUT statement. Do this only if it is actually safe to overwrite a file with data that might not yet have been loaded into Snowflake.
Note that if your Snowflake account is hosted on Google Cloud Platform, PUT statements do not recognize when the OVERWRITE parameter is set to TRUE. A PUT operation always overwrites any existing files in the target stage with the local files you are uploading.
The following clients support the OVERWRITE option for Snowflake accounts hosted on Amazon Web Services or Microsoft Azure:
Snowflake ODBC Driver
Snowflake JDBC Driver
Snowflake Connector for Python
Supported values: TRUE, FALSE.
The command cannot be executed from the Worksheets page in the Snowflake web interface; instead, use the SnowSQL client to upload data files, or check the documentation for a specific Snowflake client to verify support for this command.
File-globbing patterns (i.e. wildcards) are supported.
The command does not create or rename files.
Uploaded files are automatically encrypted with 128-bit or 256-bit keys. The CLIENT_ENCRYPTION_KEY_SIZE account parameter specifies the size key used to encrypt the files.
The command ignores any duplicate files you attempt to upload to the same stage. A duplicate file is an unmodified file with the same name as an already-staged file.
To overwrite an already-staged file, you must modify the file you are uploading so that its contents are different from the staged file, which results in a new checksum for the newly-staged file.
For security reasons, the command times out after a set period of time. This can occur when loading large, uncompressed data files. To avoid timeout issues, we recommend compressing large data files
using one of the supported compression types before uploading the files. Then, specify the compression type for the files using the
You can also consider increasing the value of the
PARALLEL option, which can help with performance when uploading large data files.
Furthermore, to take advantage of parallel operations when loading data into tables (using the COPY INTO <table> command), we recommend using data files ranging in size from roughly 100 to 250 MB compressed. If your data files are larger, consider using a third-party tool to split them into smaller files before compressing and uploading them.
Upload a file named
mydata.csv in the
/tmp/data directory (in a Linux or macOS environment) to an internal stage named
PUT file:///tmp/data/mydata.csv @my_int_stage;
Upload a file named
orders_001.csv in the
/tmp/data directory (in a Linux or macOS environment) to the stage for the
orderstiny_ext table, with automatic data compression
PUT file:///tmp/data/orders_001.csv @%orderstiny_ext AUTO_COMPRESS=FALSE;
Same example as above, but using wildcard characters in the filename to upload multiple files:
PUT file:///tmp/data/orders_*01.csv @%orderstiny_ext AUTO_COMPRESS=FALSE;
Upload a file named
mydata.csv in the
C:\temp\data directory (in a Windows environment) to the stage for the current user, with automatic data compression enabled:
PUT file://C:\temp\data\mydata.csv @~ AUTO_COMPRESS=TRUE;