What would be the preferred way of adding a canonical ref (

<link rel="canonical" href="<https://website.xyz>" />

) to an Evidence project? I've added this to the

+layout.svelte

file and am not sure if that is enough

<svelte:head>
  <link rel="canonical" href="<https://pulse.filecoindataportal.xyz>" />
</svelte:head>

Has anyone successfully deployed their Evidence project to Coolify (https://coolify.io/)? If so, how?

Hi all, My dev setup is connecting to its source via an SSH tunnel, connecting on host 127.0.0.1 When I deploy to Evidence Cloud I'll need to put in the true host address. Is there a way to have different host addresses for dev and production? Thanks

hello! if im using a data table with groutype=section and everything is grouped by a flag image, and I need subtotals how can I avoid this. I dont need the subtotal to be that string. I tried totalFmt a custom format, totalAgg a custom string, even using alt without success

@bright-smartphone-11251 - sorry to be a pain again. I'm experiencing some odd behaviour in Evidence. I had a fully working site, no problems. Then I added a new source query - next to all the other inside the source's directory. Running

npm run sources

added the parquet file to .evidence/template/static/data/[source]/[query]/ However, when I ran

npm run dev

the displayed site said it couldn't find that table. I double-checked the names. So, seeing comments on another thread with a similar problem, I purged the .evidence directory and ran

npm run sources

again. The .evidence directory builds OK, with all the parquet files as expected. When I now try to run

npm run dev

I get an error:

Unable to load source manifest

. Followed by:

No sources found, execute "npm run sources" to generate

In the app in the browser, I now get a console error:

data/manifest.json:1 Failed to load resource: the server responded with a status of 404 (Not Found)

I have updated the Evidence app, with no effect on this issue. Any thoughts? I'm at a bit of a loss here. Is there another way to rebuild the Evidence app? Possible related issue: https://github.com/evidence-dev/evidence/issues/2577 reported by @nice-jewelry-64245

OK - I've fixed the missing manifest. It turns out

npm run sources

wasn't finishing properly due to a no-longer-working test source. A previous symptom is continuing though. I thought it was related to the missing manifest... This is how the page now displays. It's like there's a CSS file missing, but I'm getting no errors in the browser console. All the icons are trying to fill the screen.

Hey team 👋 I started seeing this error with MotherDuck today:

Initialization function "motherduck_init" from file "/home/runner/.duckdb/extensions/v1.0.0/linux_amd64/motherduck.duckdb_extension" threw an exception: "

Your DuckDB version (v1.0.0) is not supported by MotherDuck. Please upgrade DuckDB to v1.2.1.

Maybe this could be fixed simply by bumping the version of

duckdb-async

used by @evidence-dev/motherduck?

Hi all! I am trying to deploy my evidence project using Cloudflare pages but I'm running into the duckdb wasm file issue where the wasm file size is too large for cloudflare pages to publish (above 25mb). I'm trying to use the r2 bucket. I've added the wasm files to the bucket, and bound the bucket to the cloudflare worker but I'm not quite sure what js to add to my project to get the wasm file from the bucket. Would really appreciate any guidance on how to resolve this. Thank you!

I have an issue with

totalAgg

for a year-over-year change column in

DataTable

.

weightedMean

doesn't work for all the cases (see attached images).

<Column id=percentage_change contentType=delta fmt=pct title="% Diff" downIsGood=True totalAgg=weightedMean weightCol=prior_year_sum />

I can get the total from the SQL query, but if I use Custom Aggregations Values to show the total, I lose the delta formatting. Is there a way to add the delta formatting to

totalFmt

?

<Column id=percentage_change contentType=delta fmt=pct title="% Diff" downIsGood=True totalAgg={yoy_mode[0].total_percentage_change} totalFmt=pct/>

how can I make the date range to be inclusive of ending? it seems to be exclusive of ending now

Quick one - i have a table i need to display pivoted. The columns are integers that need ordering in ascension - explicitly casting them has ensured this (as otherwise some part of the library is turning them to strings). The groups are cohort months - this needs to be the leftmost column. How can I get it to be without having to state every possible pivoted column name in the DataTable component: I have tried a few things like reordering the select (but it seems DataTable overrides this). I tried some funky syntax using * in the DataTable component which it finds meaningless. Any help much appreciated 🙂 - Thanks a lot!

# {params.company_type}

```sql revenue_by_company_type
select
    cohort_month,
    * exclude (cohort_month)
from (
    pivot (
        select cohort_month, cast(period_number as int) as period_number, total
        from bigquery.revenue_cohorted_by_company_type
        where company_type = '${params.company_type}'
        order by cohort_month, period_number
    ) on period_number
    using sum(total)
    group by cohort_month
)

DataTable data={revenue_by_company_type} Column id=cohort_month align=left / Column id=* align=right / /DataTable```

How to add loading indicator before the chart loads? Sorry if this question sounds dumb. I have no front end experience.

I have stacked bar charts sql filtered based on Slug parameters. The color of stacked bar chart is not consistent with all the filters though the value are same. Is there a way I can enforce it?

Hi all! Maybe a silly question, but I am trying to use BigValue and would like to show a comparison, such as X% higher vs last month. However, it is not showing for me, what am I missing?

<DateRange
    name="selected_date"
    defaultValue={'Last 7 Days'}
/>

```sql daily_spending
SELECT
    dcf.date_day,
    dcf.daily_outflow AS amount_spent,
    dcf.outflow_transaction_count AS transaction_count
FROM postgres_data.daily_cash_flow dcf
WHERE
    dcf.date_day >= '${inputs.selected_date.start}'::date
    AND dcf.date_day <= '${inputs.selected_date.end}'::date
ORDER BY dcf.date_day DESC

sql daily_average_spending
SELECT
    AVG(daily_outflow) AS average_amount
FROM postgres_data.daily_cash_flow

sql total_spending
SELECT
    SUM(dcf.daily_outflow) AS total_amount_spent,
    SUM(dcf.outflow_transaction_count) AS total_transactions,
    COUNT(DISTINCT dcf.date_day) AS days_in_range
FROM postgres_data.daily_cash_flow dcf
WHERE
    dcf.date_day >= '${inputs.selected_date.start}'::date
    AND dcf.date_day <= '${inputs.selected_date.end}'::date

<BigValue data={daily_spending} value=amount_spent format="$.2f" title={inputs.selected_date.start === inputs.selected_date.end ?

Spending on ${inputs.selected_date.start}

:

Spending from ${inputs.selected_date.start} to ${inputs.selected_date.end}

} comparison={daily_average_spending.average_amount} comparisonFormat="$.2f" comparisonTitle="vs Daily Average" comparisonDelta={true} invertDelta={true} sparkline=date_day />```

how can i update to the new version of DuckDB?

Hi.. Even though I have enabled Slash Commands in the Evidence extension in Cursor, it now has stopped working (it doesn’t show any suggestions when I e.g. wite

/Da

Yesterday I got it working again by uninstalling the extension and installing it again. But today that doesn’t work. Does anyone else experienced this? 🙏

I have a duckdb/sql question. Let me know if it's off topic. Does anyone know how to work with API that paginate? I can write a stack of UNION statements, but there must be a better way

How can we modify the data table appearance - does it conform to echart options, I can't seem to get it to: https://docs.evidence.dev/components/charts/echarts-options/#print-echarts-config Also how can I more extensively modify the appearance - for example make the sidebar togglable instead of hidden or shown etc. - any guidelines on how the styling works, if I could maybe have my own css file or something and how to incorporate that into how the page displays? https://docs.evidence.dev/reference/layouts/

During the build process, is there a way to retry failed data sources?

Hello! I am having an issue with date values output. All dates seem to be presented as d-1 instead of the actual d, where d is the actual date. For example, today is 2025-03-13, and yet when I query

current_date

in my script it shows 2025-03-12 (one day before) Is this expected? Am I missing something related to timezone or configuration here?

Hi, this query works from the duckdb cli:

select * from '<hf://datasets/fredguth/cgu__notas_fiscais/nfe_eventos.parquet>';

but this doesn’t work in evidence:

```sql eventos
select * from '<hf://datasets/fredguth/cgu__notas_fiscais/nfe_eventos.parquet>'

``` This is the error message:

IO Error: No files found that match the pattern "<hf://datasets/fredguth/cgu__notas_fiscais/nfe_eventos.parquet>"

I don’t get why.

Hello everyone, I'm currently evaluating evidence.dev for our company. I've already built a Single Page App that presents the required functionality. So far I am really impressed by evidence.dev. Now I'm having kind of a data protection issue. Using the developer tools in the browser and the Network tab, I can easily download and directly process the *.parquet files. Are there any settings I can adjust during the build process to prevent this? I filter the data in the SPA according to user profiles and don't want everyone to be able to read everything by downloading it.

👋 trying to get the evidence cli installed... been banging my head on the desk. im sure its something simple

npm run sources

> my-evidence-project@0.0.1 sources
> evidence sources

- Loading plugins & sources
✖ Loading plugins & sources
[ ! ] Cannot destructure property 'plugins' of 'getEvidenceConfig(...)' as it is undefined.

shell returned 1

I have to say that setting evidence up is not easy. The directions are so unclear. I have been going back and forth on this for a month now. I just changed from SQL server to MySQL because I could not link it up. MySQL links up but now the github directions make no sense. You state "Set up Environmental Variables" and then copy them from the button. In GitHub it has no place for that to be pasted and i have no idea why it is making me use an existing template and I still cant get it to work. this is the menu I get.

Hi all! I have this issue where there is a discrepancy with the dates (which is of date type) between the evidence query result and the result I receive using DBeaver. The dates in the raw data is "2024-02-20", however the query result for evidence returns the day minus 1, am I missing something? 😞 My timezone is Europe/Amsterdam if that matters

hey! - i have these series in a line chart, and i want to assign colors to the "Market" Series is a concat of "Market - Date..." how can I assign colors to groups? i believe this can prob be done through some js?

First of all, thank you for creating such an amazing product! Evidence has been incredibly helpful for building data-driven applications, and the documentation is very well-organized. I have a question regarding the Date Range component. Is it possible to add aliases or labels to

presetRanges

? For example: "Last 7 Days" → "過去7日" (Japanese translation). I’m translating this question into English, so I apologize if there are any mistakes or if the phrasing is unclear. Thank you in advance for your support!

Hello. When I add a BarChart to a tab, everything is displayed correctly in the app. However, when I print PDF, the BarChart is empty. BarCharts defined outside of the tabs are displayed correctly in the PDF as well. Some ideas?

How can I increase the verbosity when connecting to a source? With the same source, I get this log:

``` user_created_change ✔ Finished, wrote 5589 rows.

user_domain ✖ Error: Unknown Error```

Running

npm run source --debug

didn't show any additional errors. Running the

user_domain

query in snowflake with the same role works just fine.

Hi, I keep my evidence project at the root, at the same root as my python project that runs the pipeline to produce the data for my evidence app. Whenever I host on netlfiy or cloudflare, even if I set the base directory to be my evidence project, it always automatically detects and installs python during the build, which seems unnecessary. Does anyone know to get around this? Thanks!