This page details changes between versions of Materialize, including:
- New features
- Major bug fixes
- Substantial API changes
For information about available versions, see our Versions page.
0.1.3 → v0.2.0 (unreleased)
- Make formatting and parsing for
double precisionnumbers more consistent with PostgreSQL. The strings
[+-]Infinityare accepted as input, to select the special not-a-number and infinity states of floating-point numbers.
- Support CSV files with headers (
CREATE SOURCE...FORMAT CSV WITH HEADER).
- Support naming columns when creating CSV sources.
0.1.2 → 0.1.3
Support Amazon Kinesis Data Stream sources.
Support the number functions
round(x: N, y: N), which round
yth digit after the decimal. (Default 0).
Support addition and subtraction between
Support the string concatenation operator,
In the event of a crash, print the stack trace to the log file, if logging to a file is enabled, as well as the standard error stream.
0.1.1 → 0.1.2
SHOW CREATE SOURCEto render the full SQL statement used to create the source, in the style of
SHOW CREATE VIEW, rather than displaying a URL that partially describes the source. The URL was a vestigial format used in
CREATE SOURCEstatements before v0.1.0.
Raise the maximum SQL statement length from approximately 8KiB to approximately 64MiB.
Attempt to automatically increase the nofile rlimit to acceptable levels, as creating multiple Kafka sources can quickly exhaust the default nofile rlimit on some platforms.
Improve CSV parsing speed by 5-6x.
0.1.0 → 0.1.1
Specifying the message name in a Protobuf-formatted source no longer requires a leading period.
Indexes on sources: You can now create and drop indexes on sources, which lets you automatically store all of a source’s data in an index. Previously, you would have to create a source, and then create a materialized view that selected all of the source’s content.