Using Apache Arrow in ojs code blocks

[eitsupi]

The following documentation suggests that Observable can handle Apache Arrow.
https://quarto.org/docs/computations/ojs.html#data-sources

In our initial example we used a FileAttachment as our data source. File attachments support many formats including CSV, TSV, JSON, SQLite, and Arrow so are a convenient way to read a dataset that has already been prepared for analysis.

So I tried using the arrow file to exchange data between the R code block and the ojs code block and was confused by the strange results I was getting.

---
title: sample
format:
  html: default
engine: knitr
---

```{r}
mtcars |>
  arrow::write_feather("data.arrow")
```

```{ojs}
data = FileAttachment("data.arrow").arrow()
```

```{ojs}
Inputs.table(data)
```

This seems to be due to the fact that the Arrow library in js does not support compressed ipc files. I got the expected results by explicitly generating uncompressed files as follows

---
title: sample
format:
  html: default
engine: knitr
---

```{r}
mtcars |>
  arrow::write_feather("data.arrow", compression = "uncompressed")
```

```{ojs}
data = FileAttachment("data.arrow").arrow()
```

```{ojs}
Inputs.table(data)
```

Note that arquero generates an error when it tries to read a compressed ipc file.

---
title: sample
format:
  html: default
engine: knitr
---

```{r}
mtcars |>
  arrow::write_feather("data.arrow")
```

```{ojs}
data = aq.loadArrow("data.arrow")
```

```{ojs}
data.view()
```

```{ojs}
import { aq, op } from '@uwdata/arquero'
```

Apache Arrow is a very promising format, so it may be worth including examples in the documentation. (I used several hours for this trial and error!)

[jjallaire]

Thanks for pointing this out! I've updated the docs accordingly: quarto-dev/quarto-web@61efed2

[eitsupi]

That's should helpful!
You can close #1845 now?