PanicException: called `Result::unwrap()` on an `Err` value: NoData(ErrString("empty container given"))

[sugizo]

step
pip install -U finalytics

code

from finalytics import Screener, Tickers

screener = Screener(
    quote_type="EQUITY",
    filters=[
        # '{"operator": "eq", "operands": ["exchange", "NMS"]}' # National Market System or NASDAQ Stock Exchange work
        # '{"operator": "eq", "operands": ["exchange", "NYSE"]}' # New York Stock Exchange not work
        # '{"operator": "eq", "operands": ["exchange", "IDX"]}' # Indonesia Stock Exchange not work
        # '{"operator": "eq", "operands": ["indexmembership", "GSPC"]}' # not work
        # '{"operator": "eq", "operands": ["indexmembership", "^GSPC"]}' # not work
        # '{"operator": "eq", "operands": ["indexmembership", "JKSE"]}' # Jakarta Stock Exchange not work
        '{"operator": "eq", "operands": ["indexmembership", "^JKSE"]}' # Jakarta Stock Exchange not work
    ],
    sort_field="intradaymarketcap",
    sort_descending=True,
    offset=0,
    size=10
)

result

PanicException                            Traceback (most recent call last)
[/tmp/ipython-input-3373525991.py](https://localhost:8080/#) in <cell line: 0>()
      1 from finalytics import Screener, Tickers
      2 
----> 3 screener = Screener(
      4     quote_type="EQUITY",
      5     filters=[

PanicException: called `Result::unwrap()` on an `Err` value: NoData(ErrString("empty container given"))

best regards

[Nnamdi-sys]

Hello @sugizo ,

Yahoo finance only supports "^dji" and "^ndx" for the index membership filter.

Also for the Exchanges, the codes you have used are not the correct ones., please refer to the file below to see the correct exchange codes:
https://github.com/Nnamdi-sys/finalytics/blob/main/rust/src/data/yahoo/screeners/keys.rs

This example should work for Jakarta Stock Exchange symbols:

`from finalytics import Screener, Tickers

screener = Screener(
quote_type="EQUITY",
filters=[
'{"operator": "eq", "operands": ["exchange", "JKT"]}' # Jakarta Stock Exchange
],
sort_field="intradaymarketcap",
sort_descending=True,
offset=0,
size=50
)

screener.symbols()`

I acknowledge that I have not implemented helper data classes or validation for the Screener Filter in python unlike in the rust library. I plan to improve this in the future and add more documentation and examples for the filters.

[sugizo]

perhaps you could take some info from here (exchange, region, etc)
https://ranaroussi.github.io/yfinance/reference/api/yfinance.EquityQuery.html#yfinance.EquityQuery

compare to link above
interested in
your json file on github
that can retrieve
the symbols of components built
for index (indices)

but
tested on code above with no succeess
e.g.

'{"operator": "eq", "operands": ["indexmembership", "GSPC"]}' # not work
'{"operator": "eq", "operands": ["indexmembership", "^GSPC"]}' # not work

thanks and best regards

[Nnamdi-sys]

Hello @sugizo ,

I have tested and found that yahoo finance only supports "^dji" and "^ndx". It doesn't seem to provide components info for any other index.

The yfinance documentation of the Screener Query is really comprehensive and you can follow that as a guide.

[sugizo]

The yfinance documentation of the Screener Query is really comprehensive and you can follow that as a guide.

had you already re checked the query result ?

imo,
the query is not accurately correct for comparation
e.g.
gt, lt, gte, lte, eq, etc

because
some time the comparation missed
so the result is incorrect

my suggestion
for screener better use
tradingview_screener
it's produce more accurate comparation as a result

picture say thousand words
query :
region == indonesia
and
exchange == jakarta stock exhange

the result
show it region == US

while the same query executed with tradingview_screener
got accurate result

it's just the info,
not the financial data,
if you want to pick base on fundamental analysis that meet with your criteria
like roa > 0, roe > 0, etc

best regards

[Nnamdi-sys]

Hello @sugizo,

Thanks a lot for the feedback, I'll look into re-implementing the screener with trading view and maybe keep the yahoo screener as an option.As this would take quite some time to implement, i recommend you check out this repo that has already implemented the trading view screener:
https://github.com/shner-elmo/TradingView-Screener

I have not added Filter Metrics to the UI yet, will try to add them later.

For now you can try to use the "Sort By" Field:

[sugizo]

thanks for recommendation

that's what i usually use
every morning (before market started)
in work day

use case narrative

1. install tradingview_screener yfinance

2. query fundamental (roa > 0, roe > 0, etc)
   and recommendation (strong_buy, buy)
   with tradingview_screener

3. tradingview_screener result stock symbols
   put as list
   and
   search the list in yfinance info
   show column
   only
   fundamental data query before, including recommendation strong_buy, buy
   can put query
   on pandas dataframe too,
   but
   chosen not
   because
   the value sometime different
   between
   tradingview_screener
   and
   yfinance,
   e.g.
   tradingview roa result = 1
   while
   yfinance roa result = 0.01

4. yfinance result then rechecked with investing
   this part still not found,
   how to get the data
   from investing.com
   with python

best regards

[Nnamdi-sys]

Hello @sugizo ,

I found a possible cause for the difference you observe.

If you reference the json file for the screeners here:
https://github.com/Nnamdi-sys/finalytics/blob/main/rust/src/data/yahoo/screeners/screeners.json

you will find that some metrics have a unit: "%"

"returnonassets.lasttwelvemonths": {
      "name": "Return on Assets (ROA) %",
      "description": "The ratio of net income to total assets of a company. This measures how effective a company is using its assets to create income.",
      "data_type": "float",
      "unit": "%"
    },

I believe this means yahoo finance expresses them as decimal fractions and you would need to convert to percentage terms yourself.

Because of this type of differences, in my screener results, I return the formatted strings instead of the raw numbers.

Can you try this query and let me know if its still observe some inaccuracy:

from finalytics import Screener
screener = Screener(
quote_type="EQUITY",
filters=[
'{"operator": "eq", "operands": ["exchange", "JKT"]}',  # Jakarta Stock Exchange
'{"operator": "gt", "operands": ["returnonassets.lasttwelvemonths", 0]}',   # ROA > 0
'{"operator": "gt", "operands": ["returnonequity.lasttwelvemonths", 0]}',   # ROE > 0      
],
sort_field="returnonassets.lasttwelvemonths",  # Sort by ROA
sort_descending=True,
offset=0,
size=100 # Select Top 100
)
screener.metrics()

[sugizo]

pardon,

already aware of that,
yfinance screener
difference with
tradingview_screener
because of
percentage

problem is not on
finalytics module
but
on yfinance query

again
picture means thousand words

like in example below
you can see that
price to book value (the one that clicked blue highlighted)
is difference so high
between
tradingview_screener
and
yfinance

this is the 1st case
but
can still tolerate
because
just want to know
what is yahoo finance analyst or yahoo AI recommendation
for that emmitance (strong_buy, buy, etc)
chosen from tradingview

the the 2nd case
that
can't be tolerate is
yfinance query it self
better use
pandas query
than
yfinance query (result is more make sense)

will create comparation soon for 2nd case (unacceptatble yfinance query result)

best regards

[atcol]

I'm getting this error via the Rust API. Here's my reproducible example:

   use finalytics::prelude::*;
   use tokio;
   use anyhow::Result;
   
   #[tokio::main]
   async fn main() -> Result<()> {
       let etfs = Screener::builder()
                    .quote_type(QuoteType::Index)
                    .add_filter(ScreenerFilter::EqStr(
                        ScreenerMetric::Index(IndexScreener::Symbol),
                        "^FTSE"
                    ))
                    .build()
                    .await.expect("ETF Screener");
   
   
       etfs.overview().show().expect("");
       etfs.metrics().await.unwrap().show().expect("");
       Ok(())
   }