Retrieve column in the record struct

[Dsaquel]

Hello, i want to retrieve a column dynamically with his value just like a metadata json object with Record<String, serde_json::Value>.

The struct in question:

#[derive(Debug, Clone, Serialize, Deserialize)]
struct Item {
    #[serde(rename(deserialize = "ID"))]
    id: usize,

    metadata: HashMap<String, Value>,
}

the data looks like:

ID	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21	22	23	24	25	26	27	28	29	30	31
3007009																											LVE	LVE	LVE	LVE	LVE
3001043	LVE	LVE	LVE	LVE	LVE	LVE	LVE	LVE	LVE	LVE	LVE	LVE	LVE	LVE						MED

I that possible in csv ?

[BurntSushi]

csv isn't a typed format, so comparing it with json in this specific regard doesn't really make sense. JSON specifically has numbers, strings, arrays, objects and so on. csv has nothing but records and fields, with no inherent nesting structure, and where all fields are just strings. Any additional encoding on top of that is generally bespoke.

You might consider just using s csv::Record. That exposes the "records are just a sequence of strings" representation. That is the spiritual equivalent of a serde_json::Value.

[Dsaquel]

Thanks for your response, i see now that serde_json::Value is not appropried with csv. But can i do a Record<String, String> to catch any column with his value or something else ?

[BurntSushi]

I don't know what Record<String, String> means. csv::StringRecord exists, but it's not generic.

If you're just looking for a HashMap<String, String> that maps column names to field values for each record, then you can do that pretty easily with Serde. Just deserialize into it:

use std::collections::HashMap;

type Record = HashMap<String, String>;

fn main() -> anyhow::Result<()> {
    let data = "\
foo,bar,quux
1,2,3
hi,bye,salutations
,,
1.5,1.6,1.7
";
    let mut rdr = csv::Reader::from_reader(data.as_bytes());
    for result in rdr.deserialize() {
        let record: Record = result?;
        dbg!(&record);
    }
    Ok(())
}

Playground: https://play.rust-lang.org/?version=stable&mode=debug&edition=2021&gist=1ee4d8229f2b5c00f4e42a5fad60caaa

An example of this can also be found in the tutorial. I'd encourage you to read it. :-)

[Dsaquel]

First thanks for your time really appreciate, i was not clear but still i want to clarify because i dont know if its an issue and im sorry if im doing this in a wrong way.

If you're just looking for a HashMap<String, String> that maps column names to field values for each record, then you can do that pretty easily with Serde.

This is actually what i want but only for capturing unrecognized fields and doing it in a derive way ref.

wanted code:

use std::collections::BTreeMap as Map;
use serde::{Deserialize, Serialize};

#[derive(Debug, Clone, Serialize, Deserialize)]
struct Item {
    #[serde(rename(deserialize = "ID"))]
    id: usize,
 
    // unrecognized fields that i want catch into others
    #[serde(flatten)]
    others: Map<String, String>,
}
// later
let mut rdr = csv::ReaderBuilder::new().has_headers(false).delimiter(b';').from_path(path)?;
let items = rdr.deserialize::<Item>().filter_map(|item| item.ok()).collect::<Vec<_>>();
println!("{:#?}", items); // []

[BurntSushi]

Well I mean if you disable headers, then I'm not sure what you expect the keys of the hash map to be...

But no, for that, you would just need to use a Vec<String> to catch unrecognized values.

If you want to map them to their headers, then I'd just suggest using a csv::StringRecord instead. Like, you're doing all this work trying to use Serde, but it looks like your use case is so simple that a little bit of code to do this manually is all you need. Use Reader::headers to build a map from field index to header name. Then parse csv::StringRecord. Parse out your ID into a usize (probably should be a u64, depending) and then you have all the rest of your fields that you can access, and your map will let you get the header name for each of them.

[Dsaquel]

I was using has_headers(false) for a different file my bad i just forgot to remove it, thank you so much for your time