Cumulative dense property lost in subsequent chained measure

[HuifangYeo]

Steps to reproduce

I have a cumulative measure with dense=True.
However, when I create perform fx conversion on the cumulative measure m["Net Cashflow"], I lost the dense property as shown below by m["_convert_flow_to_usd"]:

Is this expected? I could overcome this by creating the following:

m["_cumulate_usd"] = tt.agg.sum(
    m["Signed_Amount (USD)"],
    scope=tt.CumulativeScope(level=l[("Time", "Time", "Second")], dense=True),
)

However, I was thinking it would be neater code wise, if I do the following to convert all the measurements to USD using a function:

def to_usd(measure_name):
    usd_name = measure_name + " (USD)"
    m[usd_name] = tt.agg.sum(
        m[measure_name] / m["FX Rate"],
        scope=tt.OriginScope(l["Settlement_Date"], l["Settlement_Currency"]),
    )

    m[usd_name].formatter = "DOUBLE[#,###.00000]"

Environment

• atoti: 0.7.2
• Python: 3.9.13
• Operating system: win32

[tibdex]

Hi Hui Fang, can you share some code to help us reproduce the issue faster (ideally stripped out of anything else but the code required to get to the problematic measure)?

[HuifangYeo]

Hi @tibdex,

please see the attached example: intraday-liquidity-cumulative.zip