Release Notes Version 4.2


Release Notes Version 4.2

Release Notes  

Version 4.2.3

Release Date: June 2017

Download PDF version

What’s new in 4.2?

International interactive maps

ThoughtSpot now supports international county-level maps for a select number of countries. Here are the five newly added countries, along with their administrative regions:

  • United Kingdom
    • County and Unitary Authority
    • Local Authority District
  • Germany
    • State
    • District
  • Sweden
    • County
    • Municipality
  • South Africa
    • State
    • District
  • France
    • Region
    • Department

You can use the Specify Geographic Configuration dialog while modeling your data, to select the specific GeoType setting.

Figure 1: Specify Geographic Configuration for international sub-nation regions

The names of international country regions are supported in English, as well as in their respective foreign languages.

Figure 2: Example United Kingdom geo area map

German and Japanese localization

The ThoughtSpot user interface is now available in German and Japanese. To enable localizations, please contact ThoughtSpot Support. The language used to display the user interface in ThoughtSpot is based on a setting in the user profile. If the language preference has not been set, the browser locale will be used.

Figure 3: ThoughtSpot in Japanese

Keywords, operators, and error messages are included in the translated material. However, formulas are not translated. Also, all metadata will retain their user inputted format.

This feature is supported on all browsers that support ThoughtSpot.

Nested Formulas

Nested formulas, or formula on formula, allow you to reference a formula within another formula.

Figure 4: formula1 nested in formula2

This feature enables encapsulation/decomposition of formula logic and minimizes duplication by allowing a formula to reuse another formula. These formulas become columns, and are even suggested within the Formula Builder. Invalid formulas, such as cyclic or inconsistent aggregation formulas, are prevented from being created.

In addition, these formulas have dependency awareness, which ensure that you do not break any dependent formulas while editing a formula with nested formulas. The system does not let you delete a formula that is nested in other formulas.

There is no limit to the number of nestings you can create.

Formula support for chasm trap schemas

Previously, ThoughtSpot didn’t allow you to create a formula that spanned across a chasm trap in a worksheet. However, now you can create a formula that involves aggregated measures coming from multiple fact tables of a chasm trap.

Figure 5: Formula: sum(Product Cost) / sum (Sales Cost)

The Formula Builder now provides column suggestions across chasm traps.

New keywords - Geospatial Search

The new keywords nearwithin, and farther than/from allow you to search with spatial filters on data contained in a table or worksheet which contains geo type columns with latitude/longitude data. This lets you search based on spherical distance. The distance is calculated as a straight-line distance (not road distance) radius using miles, kilometers, or meters.

Figure 6: Geospatial search on a geo area map

The keywords follow the following formats:

  • near (with default of 10 km)
  • within X miles/kilometers/meters
  • farther than X miles/kilometers/meters from

For example, your search might look like this:

revenue store name county near alameda within 50 miles

Or it might look like this:

Average hours worked branch farther than 80 km from scarborough

Case configuration

The case sensitivity for source data strings is now preserved in the display. So, the visual display of results is identical to the input case that was loaded.

It is important to note that string casings aren’t applied globally, but by column. So datasets will have different string casings as long as they’re in different columns. Tables that are already compacted will keep their lowercase format. In these cases, in order to get the specific string case that you want, you would have to truncate all related tables and reload them.

To take advantage of case configuration, you need to have ThoughtSpot Support enable it on your cluster. In addition, title casing should be disabled for string casing to properly work.

Freeze column capability for data modeling page

As you edit the column properties of a table or worksheet, the Column Name column now freezes in place. This should make it much easier for you to know which column you’re making changes for.

Data Profile

Basic column data profile information is available under Profile on the Data page.

Figure 7: Column profile view

When data is first loaded into ThoughtSpot, it is often unfamiliar to users. This profile view should help you get a better sense of what’s there before asking questions on the data.

Expanded Joins for Aggregated Worksheet

You can now join aggregated worksheets to any base table or user imported data. In addition, aggregated worksheets can be included as tables in worksheets. This means aggregated worksheets now behave similar to tables, and can be used in the same way as a table, excluding any TQL manipulation.

Inclusion Rule (Left Outer Join | Inner Join | Full Outer Join)

ThoughtSpot now supports full outer and right outer joins during worksheet creation, regardless of the sharding of underlying tables.

Figure 8: Inclusion rule choices

Non co-sharded joins

ThoughtSpot also supports joins between two sharded fact tables with different shard keys, otherwise known as non co-sharded tables. However, because data is being redistributed when this type of join happens, search results will come back more slowly than with co-sharded tables. Therefore, we recommend that you use a common shard key for two fact tables.

You are not limited by the column connection or relationship type.

TQL Dependency Checker

TQL prevents disruptive changes by checking for dependencies when schema changes are made. In the case of a potential breakage, you are shown the dependent objects.

Figure 9: Dependency awareness in TQL

Dictionary based query optimizations

Dictionary compression, which replaces the original data with a map string, is already implemented in your ThoughtSpot application. This reduces memory usage, as the system only has to look at the column value associated with the original data, then look that up in the compressed dictionary to see what it equals to.

Now, we are leveraging dictionary compression to speed up queries, especially those involving strings. Query executions should only take place on the compressed dictionary domain, where only a small number of values have to be decompressed before returning an answer. This eliminates decompression overhead, and should improve queries that involve strings.

Automatic periodic snapshot

There is a new implementation of snapshots—manual or periodic. So you are shown two lists when you run the tscli command.

You can have up to 20 manual snapshots, after which, you have to clear one before you are able to create another.

Automatic snapshots follow time-based policies, e.g.:

  • 6 * 10-minutes snapshots
  • 3 * 1-hour snapshots
  • 2 * 4-hour snapshots

These policies are configurable.

Automatic periodic backup

Time-based automatic backups are available as well, with an option of dataless or full backups. You can also take dataless backups locally.

There is also a richer garbage collection policy e.g.:

  • 7 * 1-day backups
  • 4* 1-week backups

And like automatic snapshot policies, backup policies are also configurable.

Upgrade management

We have made upgrades more robust and controllable in the following ways:

  • Preupdate checks for potential failures
  • Ability to check the status of the update process
  • Resumable upgrades from where it paused
  • Bypass failures to continue upgrades

Support for large number of users on an appliance

ThoughtSpot can support up to 600 simultaneous users, and provide scale for more than 80,000 users in the system.

ThoughtSpot Data Connect improvements

There are a number of improvements around the self-serve data integration UI offered through ThoughtSpot.

Data Connect data source re-configuration

You can add or remove columns for existing tables. This means you can edit the data source definition to add or remove columns, and adapt existing jobs to change source schemas.

Figure 10: Re-configuring the data source

All relationships are captured with this feature. The relationships are automatically created for you when you add a primary key column.

It is recommended that you do not edit the schema when altering tables.

Data Connect script-based processing

You now have to option to process the data before and after any load by adding a pre- or post-script. A script is a set of TQL statements, and you can add a script of any TQL command. The main use case is to truncate existing tables.

Figure 11: Options to add pre- or post-scripts

Other UX improvements

Other UX improvements include chart and geo map improvements.

Enhanced color scheme for charts

Colors are maintained across searches within a session. For example, when doing a search on revenue by state, each state will keep its color assigned to it even if you change the search or chart type.

Enhanced configurator

The chart configuration panel has been combined and moved to the right drawer on the answer page. This new configurator includes the option to manually set the y-axis scale.

Figure 12: The enhanced chart configurator

Pareto chart enhancement

The line portion of the pareto chart has been corrected to show data labels as percentages.

Geo map data labels

You can now see data labels on geo area and geo bubble maps by checking the Show Data Labels setting in the chart configurator.

Color scale on geo maps

You can now see color scale on the bottom right corner of the map for the geo area chart. The color scale will only show up when there is no legend column, or the legend column has only one value selected.

Background filled world map

When geo tiles are not enabled, we now show a filled world map in the background to give you a sense of the current location and zoom level on the map. The maps will show water as white and land as light gray.

Faster loading of bubble and heatmap

We now fetch only points, instead of the whole polygon boundaries, to draw bubble and heat maps, which makes them load faster.

Unsupported zip code area map

Geo area maps are no longer supported for zip codes. This is because zip codes are not comprised of areas, and it is therefore impossible to get the boundary data for zip codes. Earlier we supported it using a data file that only had boundaries for half of the zip codes.

New Features in 4.2.3

Embed: ThoughtSpot Navigation bar 

Beginning with the 4.2.3 (hotfix) users must use a combination of the embedApp and primaryNavHidden flags to configure the visibility of the ThoughtSpot Navigation bar.  To hide the primary navigation, configure this:

  • Make sure the app is in an <iframe/>.
  • Set the embedApp flag as true. This flag determines if the application is embedded.
  • Set the primaryNavHidden flag as true (the default). This flag determines navigation visibility.

If either flag is false the primary navigation will appear.  

Ability to Change the Data Connect admin credentials

Data Connect creates a ThoughtSpot connector on Informatica Cloud. This connector ensures communication between Informatica and ThoughtSpot is secure. The ThoughtSpot connector uses a single administrative user to connect back to ThoughtSpot Data Connect.  You provide this admin user’s credentials when you enable Data Connect. Typically the user is tsadmin but on your installation it may be different.

In this release, the ability to change the credentials for the ThoughtSpot connector administrative user was added. To change the credentials, do the following:

  1. Disable Data Connect.

     tscli etl disable-lw

  1. Enable Data Connect  again but provide new credentials when you specify the --admin_username flag.

      tscli etl enable-lw ...

Data sources created under the previous credentials are accessed now by the new ones.

Option added to users/sync API

The removeDeleted parameters was added to the /tspublic/v1/user/sync API.  This is a boolean flag that indicates whether to remove deleted users/groups. When true, this flag removes any deleted users or groups.

Ability to disable the help carousel pages in embed mode

You can use the tscli command to disable the Search Doctor help feature which includes the tips, tricks, video, and  carousel.  This is intended for sites using ThoughtSpot in embed mode. Speak with your ThoughtSpot Customer Service representative to learn how to do this.  

Ability to filter by creator on Pinboards, Answers, and Data pages

Users now have the ability to search by creator name on the Pinboards, Answers, and Data pages.

What’s fixed in 4.2.3?

FIXED: Previous successful JOIN conditions produced incorrect  


Several JOIN conditions failed to work as they had in a prior release. This has been fixed.

FIXED: Unauthorized users had access to Data Sources tab


Description: Users without Data Management privileges could see the Data Sources tab and add new sources.  This is fixed.  The tab is greyed-out and users are instructed to contact their administrator for authorization.

FIXED: Complex queries with unique count timed out

ISSUE ID: SCAL-17975, SCAL-17929

Description: Complex queries with unique count caused the system to timeout. This occurs most frequently when queries involve chasm-trap type data.  This has been fixed.

What’s fixed in 4.2.2?

FIXED: Horizontal scrolling in a pinboard not working well


Description: Trying to scroll horizontally (towards the right) on a pivot table within a pinboard, caused the columns to scroll as expected, but the very first column on the left started moving towards the right as well, taking the whole window along with it. This has been fixed.

FIXED: Download pinboard not working


Description: Download pinboard wasn’t working in version 3.5.7. This has been fixed.

FIXED: Overlapping labels in pivot table on IE11


Description: In the Internet Explorer 11 browser the label for a pivot table column had overlapping letters. This has been fixed.

FIXED: “Show underlying data” doesn’t show anything


Description: When using a Safari browser, choosing “Show underlying data” from a pinboard showed a message saying that some rows are being shown, when in fact no rows are shown. This has been fixed.

FIXED: Cannot see users with access to pinboards


Description: When editing multiple pinboards in bulk, to share with users and groups, everything works as expected. But afterwards, when editing permissions on the pinboards, only the groups with access, and not the users, are shown. However, the users still have permission to access the pinboards. This has been fixed, so that you now may view and edit the users who have access.

FIXED: Error shown instead of placeholder title for formula


Description: On the Internet Explorer 11 browser, trying to edit a formula showed an error message where the formula title should be. This has been fixed.

FIXED: Context menu to drill down from a pinboard is not available


Description: A user with View access to a pinboard and with access to the underlying data was unable to access the drill down menu on the pinboard when viewing, editing, or presenting. This has been fixed.

FIXED: Cannot disable NPS survey question


Description: After using the flag to disable the Net Promoter Score (NPS) question at the bottom of the browser window, the NPS question was still showing. This has been fixed.

FIXED: Fiscal year-to-date formulas not working


Description: Formulas that contained fiscal year-to-date calculations were not working previously. This has been fixed.

FIXED: Filtering on null values causes no data to display


Description: In the filter dialog, null values were displaying as {null} instead of {blank}. When selecting to filter on these values, the message “No data to display” was shown. This has been fixed.

FIXED: Remove external join paths from worksheet scope


Description: When creating a worksheet, you are prompted to select a join path if there are multiple possible join paths. In the past, all possible join paths were shown, including those that involved tables outside of the worksheet scope. This made it hard to choose an appropriate join path. Now only join paths within the scope of the worksheet will be shown.

FIXED: Could not edit users in pinboard sharing


Description: When editing a group of shared pinboards, the groups with permissions on the pinboards could be changed, but the users with permissions didn’t appear. This has been fixed.

FIXED: Using IS NULL when defining a generic relationship not working


Description: Using a generic relationship which included the IS NULL operator was causing the application server to fail. This has been fixed.

FIXED: Aggregated worksheet data values not showing in search suggestions


Description: For aggregated worksheets, data values were not being indexed for search suggestions. Now, if the value is an attribute with aggregation set to “none”, the value from the base table will be used to generate the search suggestion for the aggregated worksheet.

FIXED: Need ability to set Attribution Dimension on a column in a worksheet

ISSUE ID: SCAL-18478, SCAL-17009

Description: In the past, you could set the Attribution Dimension column setting on a base table, but not on a worksheet. Some use cases required this setting to be different for different worksheets built on the same base tables. Now you can make this setting at the worksheet level.

FIXED: Non cosharded join caused system to hang


Description: A worksheet that required joining tables that were not cosharded caused the system to hang. Checks have been put in place to prevent this in the future.

FIXED: Error when uploading modeling file


Description: After downloading and editing the modeling CSV file, uploading it produced an error. This has been fixed.

FIXED: Scheduled pinboards are not being sent out


Description: Scheduled pinboards email messages were not being sent out. This was due to a discrepancy between the capitalization in the recipients email addresses. This has been fixed.

FIXED: Add the ability to disable/enable localizations


Description: You can now enable and disable locale based translations of the ThoughtSpot application. These are disabled by default. If you want to enable or disable localizations, please contact ThoughtSpot Support for assistance.

FIXED: Lock contention during long running queries


Description: Long running searches were blocking answers from being returned for searches issued after the long running search. This has been fixed, so that those answers which can be returned will be, even if another search is taking a long time.

FIXED: Add the ability to protect a snapshot from deletion


Description: Added the ability to “pin” a snapshot so that it doesn't get automatically garbage collected. This is useful in cases where you want to save an automatic snapshot longer than it would normally be saved. For details on how to accomplish the pinning or unpinning of a snapshot, contact ThoughtSpot Support.

FIXED: Corrected pinboard impressions and ad hoc search metrics

ISSUE ID: SCAL-14765, SCAL-18432, SCAL-17841, SCAL-17970

Description: The pinboard impressions and ad hoc search metrics were computed in such a way that they undercounted impressions. This has been fixed.

FIXED: System tables getting indexed for search incorrectly


Description: Some new system tables were being indexed for search by default, when they should have been set to not be indexed. This has been fixed.



FIXED: Active Directory sync created duplicate users

ISSUE ID: SCAL-19122, SCAL-18842

Description: Active Directory sync was creating users in ThoughtSpot based on their memberships in AD groups. But the domain name was being created in lowercase (even when AD had the domain name in uppercase). This was causing duplicate users to be created. This has been fixed.

FIXED: SAML auto redirect causes page not found error


Description: The SAML auto redirect was causing a HTTP 404 Page Not Found error. This has been fixed.

FIXED: Security: Strong SSL security in nginx


Description: Implemented the latest security recommendations for nginx, including:

  • Only enabling TLS 1.2; disabling TLS 1.1 and 1.0

  • Disallowing unsafe ciphers

  • Added a setting to enable choosing the server’s cipher preference, rather than defaulting to the client’s preference

FIXED: Data Connect: Provide a way to specify proxy details when enabling Data Connect


Description: When enabling the infaagent for Data Connect with a proxy, the setting was not persistent over a software update. You had to manually re-set the proxy for Data Connect. This has been fixed.

FIXED: Data Connect: Not able to edit data sources when target tables include generic relationships


Description: When the target tables in ThoughtSpot included generic relationships (those created using the syntax “CREATE RELATIONSHIP WITH”), the Data Connect data sources could not be edited. This has been fixed.



FIXED: Data Connect: Informatica agent infraagent does not come up upon restart


Description: When restarting ThoughtSpot, the infraagent did not restart, even though it was reported as up and running. This has been fixed.

FIXED: Data Connect: Loads in progress could not be restarted or aborted


Description: Some loading tasks that were in progress didn’t have the ability to be stopped or  restarted. This was due to the call to the Informatica API login failing or timing out. This has been fixed.

FIXED: Data Connect: Tasks failing due to timeout


Description: Timeout was causing Data Connect tasks to fail when fetching a table. This has been fixed.

FIXED: Data Connect: Notification of failed job didn’t go out if Informatica job was deleted


Description: After deleting tasks on Informatica, email alerts were not sent out when the tasks didn't start due to these tasks not being found. This has been fixed.

FIXED: Extended Enterprise Edition: Add custom branding to the browser tab favicon and text


Description: In ThoughtSpot Enterprise Edition, you can now modify the browser tab favicon and text to reflect your company’s brand. When you upload a new application logo, it will also be used as the favicon for the browser tab. You can also enter custom text that will appear as the page title in the browser tab. To enter custom text, go to the Page Title section at the bottom of the Style Customization page and enter the new text there.

Figure 13: Customizing the favicon for the browser tab

FIXED: Extended Enterprise Edition: Remove Copy Link function from embedded visualization


Description: The Copy Link icon should have been hidden from embedded visualizations, because its function is to provide the link to use for embedding. It has been removed.

What’s fixed in 4.2.1?

FIXED: Record client's identifier in statsDB


Description: You can now find out the user id and session id for API calls. This helps you track stats like who the daily active users are.

FIXED: Login screen can be redirected to SAML location


Description: You can have your users skip the ThoughtSpot login screen and be immediately redirected to your SAML location.

FIXED: tscli to increase number of points on a geo chart


Description: Before you could only use a URL flag to increase the data points shown on a geo chart. Now there is a tscli command to increase the count to over 5,000. Contact ThoughtSpot Support if you would like to adjust this limit.

FIXED: ThoughtSpot doesn’t work on Firefox and Safari browsers

ISSUE ID: SCAL-18132, 18405, 18551

Description: The pinboard page didn’t load when using ThoughtSpot on Firefox, among other issues. ThoughtSpot now works on Firefox and Safari version 10.

FIXED: Error - "object could not be opened due to errors in some of its dependencies"


Description: Opening a worksheet with no missing dependencies would return an error. This would also affect row level security.

FIXED: Periodic backup failure due to nodes not mounting NAS


Description: A periodic backup failure would occur when there were four nodes that have not mounted NAS due to permissions on the NAS server side. In the future, NAS should be mounted correctly for all nodes using the tscli nas mount command before implementing periodic backup.

FIXED: Schema Viewer doesn't work with localization


Description: Clicking View Schema in the Data view returned a blank screen when using a localized version of ThoughtSpot. This has been fixed.

FIXED: SAML auto redirect logout returns empty page


Description: The automatic SAML logout redirect was broken, and would return an empty page.

FIXED: /session/user/createusers doesn't create users in bulk


Description: The /session/user/createusers endpoint has changed and would not create users in bulk. This has been fixed.

FIXED: Data Connect - Cannot completely remove data sources


Description: Deleting a data source would not completely remove it from the UI. Instead, an error would occur and the data source would persist.

FIXED: Data Connect - Editing a data source changes metadata that breaks underlying objects


Description: Editing a data source that references an existing table would change the metadata. This would then break underlying objects.

What’s fixed in 4.2?

FIXED: Aggregation on top of an aggregated formula error


Description: An error message was returned when you performed an aggregation, such as sum, on top of an aggregated formula. This has been fixed by preventing you from saving inconsistent aggregations.

FIXED: Can’t add new filter if there’s a blank filter


Description: You could not add any additional filters to a pinboard if you already had a blank filter associated with it.

FIXED: Pinboard slideshow resets after resetting drill down


Description: While presenting a pinboard, drilling down then resetting would take you back to the beginning of the slideshow. Now, the slideshow remains on the visualization being resetted.

FIXED: Case sensitivity is not preserved when importing URLs


Description: Case sensitivity was not preserved when importing URLs to the data view. This issue is important since some URLs require case sensitivity. This has been fixed by support for mixed case in output.

FIXED: Group_function ratios are not summing to 100%


Description: Group_function ratios should sum to 100%, but this is not the case due to a rounding issue.

FIXED: Share dialog causes line wrapping for users or groups with very long names


Description: The share dialog caused line wrapping when users or groups with very long names were added to the share queue. This caused confusion as to which users or groups had been selected.

FIXED: Unable to sync users/groups from AD using LDAP sync script


Description: The users that were locally created in ThoughtSpot were being synced with AD but were not being dropped into their corresponding AD groups. This has been fixed.

FIXED: Various data label issues with column and line charts


Description: There were various issues with data labels and the tooltip when negative values or shared y-axis was involved. This has been fixed.

FIXED: Show underlying data broken for no_bucket/ detailed timestamp field


Description: This error only occurred when one of the fields was a timestamp only field, and the bucket for it was set to no_bucket/detailed. When this field was removed, the underlying data worked.

FIXED: Rename Column in TQL makes it unalterable


Description: When you changed the name of a column on a base table using the TQL shell you could no longer use that column in other alter table statements. This issue has been fixed.

FIXED: Firefox/IE: Pinboard - Scrollbars inside the pinned visualizations are unresponsive to dragging


Description: The scrollbars on the visualizations in pinboards would not properly respond to dragging. Instead, the visualization would follow the cursor around. This stuck-mouse bug has been fixed.

FIXED: Unable to determine which table is which


Description: If you had multiple tables with the same or similar name in different databases you might have had a difficult time telling them apart. For example:

  • vwBridge_DimSecuredDWUser_DimAccount
  • vwBridge_DimSecuredDWUser_DimAuditForm
  • vwBridge_DimSecuredDWUser_DimAuditType

This issue has been fixed by allowing the display to show the database/schema along with the tables.

FIXED: Unable to open downloaded Excel

ISSUE ID: SCAL-16854, 17388

Description: Downloaded excel files with long names would not open. This was because the exported filename exceeded the limit set by Microsoft Excel. This has been fixed.

FIXED: Sage suggestions and formula definition are not rendered on IE 11

ISSUE ID: SCAL-16931, 16932

Description: While using Internet Explorer 11, editing a token in the middle of a query provided an option to scroll down for other suggestions, which is not the correct behavior. Also, editing a formula in an unsaved answer would render the formula editor with random text and a perpetual loading indicator.

FIXED: Download as CSV throws exception


Description: This issue was related to how answers for visualizations were generated when a pinboard filter was applied. This has been fixed.

FIXED: Row data still being shown in filters on worksheet with chasm trap involved


Description: Data that shouldn’t be seen was shown in filters on worksheets with join paths on tables involved with chasm traps. This regression bug has been fixed.

FIXED: Remove deprecated pinboards/worksheets


Description: We have removed deprecated pinboards and worksheets that are no longer being used on the Resource Usage page.

FIXED: Chart legend labels are cut off


Description: Chart legend labels were cut off even when there was plenty of room.

FIXED: CSV/Excel downloads now indicate when results are incomplete


Description: CSV and Excel files now indicate at the top of the page the number of rows that have been successfully downloaded.

FIXED: SAML login link not appearing


Description: The SAML login link would not appear on the login page. This blocked SAML use and portal integration. The temporary workaround was to use the following link: http(s):///callosum/v1/saml/login

FIXED: Asynchronous rendering of axis and data labels


Description: There were significant delays in load time for pinboards due to synchronous rendering of all labels. Deciding to load labels asynchronously has reduced the load time.

FIXED: Format pattern disabled for non-additive attributes


Description: Format pattern option in data modelling was disabled for non-additive numeric attributes. This was an issue with the UI.

FIXED: Filtering on day_of_week function fails


Description: Filtering on day_of_week function from the search bar failed and returned a red bar error.

FIXED: Unable to modify fields via model file



Problem: Field descriptions for tables and worksheets were broken. This was due to the metadata cache not being cleared upon the updates.

FIXED: Incorrect counting of nodes


Description: This was due to a bug in, whereby a new node would get counted when we tried to count the number of old nodes. This only happened when we tried to re-add an old node.

FIXED: Synonyms are not suggested in search bar


Description: Synonyms were not being suggested in the search bar. This was because data value suggestions were over prioritized over synonym suggestions.

FIXED: Cluster doesn’t start automatically after one node cluster booted

ISSUE ID: SCAL-17976, 17998

Description: It seemed that one node was waiting on some kind of process when the process had not started and it was not listening on a specific port. The workaround was to manually stop and start the cluster.

FIXED: Pinboard PDF download sometimes contains individual visualizations that have not finished rendering


Description: Some visualizations would not be completed rendered in a PDF download.

FIXED: Data Connect - Tables imported from data connect don’t have column level security option


Description: Data Connect imported tables need to be just like regular tables and support column level security. Row level security can now be applied to tables imported via Data Connect, and the RLS rules are correctly honored.

FIXED: Data Connect - Getting error when setting up Data Connect


Description: We would get a “No Json could be decoded” error when setting up Data Connect. This issue came about because the infa cloud IP addresses were not whitelisted.

FIXED: Data Connect - Suggestions for column values are not appearing in the search bar


Description: Suggestions for column values were not appearing in the search bar. This was because special characters prevented filtered results.

FIXED: Data Connect - Deleting a user removes jobs created by the user


Description: Deleting a user deleted all Data Connect jobs created by that user. Now, when a user is deleted, the unclaimed jobs are reassigned to tsadmin.

FIXED: Data Connect - Falcon column changes don’t propagate to Calllosum


Description: You can now rename columns and tables after a data load, using the DDL edit step or the ThoughtSpot web application.

FIXED: Data Connect - ($LastRunTime) - Set to GMT only


Description: The $LastRunTime was permanently set to GMT timezone and couldn’t be reconfigured. You are now able to add an expression on top of $LastRunTime/$LastRunDate for a filter condition.

FIXED: Data Connect - Data load failed when initializing writer due to SSL cert


Description: An error would appear stating that the data load failed when initializing writer due to SSL cert. This has been fixed.

FIXED: Data Connect - data load failing for date type column with JDBC


Description: With Data Connect, data load for a table was failing for date type column with JDBC connector. This error occured while writing data to the target native system.

Known issues

Applying filter on {blank} value results in no data


Description: Applying a filter on {blank} values results in an answer with no data.

Proper error message for tsload errors


Description: Error messages when loading data using tsload are incomplete.

tscli monitoring command should check validity of email address


Description: tscli monitoring set-config --email does not check the validity of email addresses. Instead, the command takes whatever value was enter.

KPI line extends outside of the chart canvas


Description: The range of the y-axis is not taking the KPI value into consideration, so the line exists beyond the scope of the chart canvas.

Resizing a column automatically sorts it


Description: When resizing the first column of a table, the action automatically sorts the column.

Show underlying data is broken on aggregated worksheets


Description: Selecting show underlying data on aggregated worksheets sometimes returns unexpected results.

Explanation for non co-sharded join prevention


Description: When you try to add a column that causes a non co-sharded join, you get a generic error message. The message should explain how the error came about.

Exclude is broken on measures


Description: Attempting to use Exclude on measures provides an irrelevant option. Exclude for measures will be replaced with “Show MEASURE >= NUMBER” and “Show MEASURE <= NUMBER”.

Single disk failures causes cluster update failures


Description: Single disk failure in a cluster derails the entire upgrade workflow, requiring manual intervention followed by a retry of the upgrade. Rebooting helps with the upgrade.

Downloaded charts are poor quality on IE 11


Description: Charts downloaded while on Internet Explorer 11 are fuzzy and difficult to read.

Aggregated worksheet data value index - point to base column index


Description: There are no data value suggestions for aggregated worksheets.

Supported upgrade paths

If you are running a release that begins with one of the following ThoughtSpot version numbers, you can upgrade to this release directly:

  • 3.5.x, 4.0, 4.1, 4.2

If you are running a different version, you need to do a multiple pass upgrade. First, upgrade to one of the above versions, and then to this version.