Issues fixed in MarkLogic Server 9.0-9.1

ODBC Performance & Debugging

Data accessed via ODBC is slow. Unable to disable ODBC debugging.

SQL result double values truncated over ODBC

SQL result double values are incorrectly truncated to six digits of precision over ODBC.

Segmentation fault

Segmentation fault in xdmp::ScratchExpandedTreeStorage::addBinaryNode

Segmentation fault in entity extraction

When the matches from a call to cts:entity-extract include matches with the same starting and ending location as well as the same keyword, a segmentation fault may occur.

PowerBI & boolean types

Booleans do not work in PowerBI

ECDH key exchange

ECDH key exchange support now available with this release

Optimizations on timestamp functions in SQL

Power BI data slicer on date column produces inefficient queries that has filters not a candidate for index scan. Rewriting the query in an optimized way will fix the query to still use index scan.

A nested for with an order by may segfault

Under some circumstances, a 'for' within a 'for' with an order by clause with multiple ordering constraints may cause a segmentation fault. The pattern is 'for... for... order by $something, $another ... return ...'

Query Console query text to request body

When saving a Query Console query to the server via an AJAX request, the content of the query was appended to the query string, rather than being sent via a POST/PUT body. This fix restores that payload to the body.

Temporal updates combined with node updates may create duplicate documents

Performing a temporal update and a non-temporal node update in the same statement may create duplicate documents.

HTTP Server mishandles HTTP requests with zero headers

The server does not return appropriate response to HTTP requests with zero headers

Integer scalar type treated as strings in PowerBI

Have PowerBI treat xs:integer as a string. This is because xs:integer can have values outside of the range of PowerBI's biggest data type, the Whole Number. See https://docs.microsoft.com/en-us/power-bi/desktop-data-types Because of this limitation, PowerBI cannot "fold" direct query expressions with xs:integer. Recommend using xs:int instead of xs:integer in PowerBI. If you need bigger numbers than xs:int can hold, recommend using xs:long. Updated xs:integer to be casted as char in the Direct Query connector unless this PowerBI limitation is addressed by Microsoft.

Complex fields with range indexes and tokenizer overrides Segfault

Given a field defined with a range index and tokenizer overrides, inserting, reindexing, refragmenting, replicating, or rebalancing a document has an instance of the field that involves multiple child elements may cause a segmentation fault in the tokenizer.

Redaction fails for JSON objects with colon in key name

When there are JSON object with colon in key name like "foo:bar", redactions in child nodes below this level are not redacted

Long casted as float in PowerBI issued queries

Long type data shouldn't be casted as a float when using the MarkLogic direct query data connector.

SQL query optimization for "is NULL" and "is not NULL" on a non-nullable column

SQL queries containing conditions “column is NULL” and “column is not NULL” on a non-nullable column will be optimized. Also when these kind of filters are used in “OR”/”AND” condition with other filters in SQL queries, then relevant optimizations will be applied to simplify the filter processing.

mlcmd won't create duplicate /etc/fstab records when the VM reboots on Azure

When deploy MarkLogic on Azure VMs, mlcmd will try to format and mount disk every time it runs. Although the formatting and mounting operation will fail but a record will be appended to /etc/fstab.

Support for client certificates that don't identify themselves

When creating a client certificate, you can now add a "certificate-type=client" parameter to the secure-credentials request to explicitly identify the certificate as a client certificate if it does not include the extension attribute that says it is.

PowerBI String/text Data Slicer optimization

The String/text Data Slicer (in Power BI) currently produces SQL that is not optimal.