Load data into PostgreSQL
pgloader is a data loading tool for PostgreSQL, using the
Its main advantage over just using
\copy, and over using a
Foreign Data Wrapper, is its transaction behaviour, where pgloader
will keep a separate file of rejected data, but continue trying to
copy good data in your database.
The default PostgreSQL behaviour is transactional, which means that any erroneous line in the input data (file or remote database) will stop the entire bulk load for the table.
pgloader also implements data reformatting, a typical example of that
being the transformation of MySQL datestamps
0000-00-00 00:00:00 to PostgreSQL
NULL value (because our calendar
never had a year zero).
pgloader version 1.x is quite old and was developed in
When faced with maintaining that code, the new emerging development
team (hi!) picked
python instead because that made sense at the
time. So pgloader version 2.x was written in python.
The current version of pgloader is the 3.x series, which is written in Common Lisp for better development flexibility, runtime performance, and support of real threading.
The versioning is now following the Emacs model, where any X.0 release
number means you're using a development version (alpha, beta, or release
candidate). The next stable versions are going to be
When using a development snapshot rather than a released version the version number includes the git hash (in its abbreviated form):
pgloader version "3.0.99"
Release candidate 9 for pgloader version 3.1, with a git tag named
v3.0.99so that it's easy to checkout the same sources as the released code.
pgloader version "3.0.fecae2c"
Development snapshot again git hash
fecae2c. It's possible to have the same sources on another setup with using the git command
git checkout fecae2c.
pgloader version "3.1.0"
pgloader is available under The PostgreSQL Licence.
$ apt-get install pgloader
You can also use a docker image for pgloader at https://hub.docker.com/r/dimitri/pgloader/:
$ docker pull dimitri/pgloader $ docker run --rm --name pgloader dimitri/pgloader:latest pgloader --version $ docker run --rm --name pgloader dimitri/pgloader:latest pgloader --help
Build from sources
When building from sources, you should always build from the current git
HEAD as it's basically the only source that is managed in a way to ensure
it builds aginst current set of dependencies versions.
Building from sources on debian
$ apt-get install sbcl unzip libsqlite3-dev make curl gawk freetds-dev libzip-dev $ cd /path/to/pgloader $ make pgloader $ ./build/bin/pgloader --help
Building from sources on RedHat/CentOS
See "Redhat / CentOS" in INSTALL.md
Building from sources on macOS
When using brew, it should be a simple
brew install --HEAD pgloader.
When using macports, then we have a situation to deal with with shared objects pgloader depends on, as reported in issue #161 at https://github.com/dimitri/pgloader/issues/161#issuecomment-201162647:
I was able to get a clean build without having to disable compression after symlinking /usr/local/lib to /opt/local/lib. Note that I did not have anything installed to /usr/local/lib so I didn't lose anything here.
Building from sources on Windows
Building pgloader on Windows is supported, thanks to Common Lisp implementations being available on that platform, and to the Common Lisp Standard for making it easy to write actually portable code.
It is recommended to have a look at the issues labelled with Windows support if you run into trouble when building pgloader:
Building Docker image from sources
You can build a Docker image from source using SBCL by default:
$ docker build .
Or Clozure CL (CCL):
$ docker build -f Dockerfile.ccl .
More options when building from source
pgloader knows how to produce a Self Contained
Binary file for pgloader, found at
$ make pgloader
$ make CL=ccl pgloader
SBCL and it supports core compression, the make process will
use it to generate a smaller binary. To force disabling core
compression, you may use:
$ make COMPRESS_CORE=no pgloader
--compress-core is unique to SBCL, so not used when
CC is different
You can also tweak the default amount of memory that the
will allow itself using when running through your data (don't ask for more
than your current RAM tho):
$ make DYNSIZE=8192 pgloader
make pgloader command when successful outputs a
file for you to use.
$ ./build/bin/pgloader --help $ ./build/bin/pgloader <file.load>
For example, for a full migration from SQLite:
$ createdb newdb $ pgloader ./test/sqlite/sqlite.db postgresql:///newdb
Or for a full migration from MySQL, including schema definition (tables, indexes, foreign keys, comments) and parallel loading of the corrected data:
$ createdb pagila $ pgloader mysql://user@localhost/sakila postgresql:///pagila