as.Date NA error and format invalid 'trim' argument

[Felipe1]

Hi guys

I cant do the format that i want to my data

as.Date(price$date, "Y%-%m-%d")

NA error

format(price$date, "Y%-%m-%d")
#invalid 'trim' argument

I have a problem with locale i guess

Other information about my data:

class(price)
[1] "data.frame"

str(price)
'data.frame': 19285 obs. of 3 variables:
crypto: Factor w/ 40 levels "ADA","ALGO","ATOM",..: 34 12 38 32 18 39 1 5 20 31 ... price : num 11194 389.8 0.3 10005 98602 ...
$ date : Factor w/ 1040 levels "01-01-2018","01-01-2019",..: 151 151 151 151 151 151 151 151 151 151 ...

---

[FJCC]

You have to tell as.Date() what format the date is in, not the final format you want. If it is currently d-m-Y, use

as.Date(price$date, "%d-%m-%Y")

[Felipe1]

Hi

I can't transform the data frame to a times series

class(price2)
str(price2)

price2$date <- as.Date(price2$date)

as.Date(price2$date, "Y%-%m-%d")

xts_price2 <- xts(price2$crypto, price2$price, price2$date)

xts_price2 <- xts(select(price2,-c("crypto","date")),
order.by = as.POSIXct(strptime(price2$date,"%Y-%m-%d")))

Errors:

Error in xts(price2$crypto, price2$price, price2$date) :
order.by requires an appropriate time-based object

xts_price2 <- xts(select(price2,-c("crypto","date")),

•               order.by = as.POSIXct(strptime(price2$date,"%Y-%m-%d")))
  

Error in xts(select(price2, -c("crypto", "date")), order.by = as.POSIXct(strptime(price2$date, :
'order.by' cannot contain 'NA', 'NaN', or 'Inf'

Here the photos of data frame

[FJCC]

Let's make sure your original problem of using as.Date is solved. If you original date column is in the format d-m-y, you need to use code like

price$date <- as.Date(price$date, "%d-%m-%Y")

Running str(price) after that should show you that the conversion to a numeric date was successful. Can you confirm that?

[Felipe1]

@FJCC

"%d-%m-%Y"

yes now is ok i guess

[FJCC]

The following command will not change the values in price$data to be dates.

as.Date(price$date, "%d-%m-%Y")

It will only display the result of that action on the console. To store the result in the price data frame, you have to do

price$date <- as.Date(price$date, "%d-%m-%Y")

[Felipe1]

@FJCC

price$date <- as.Date(price$date, "%d-%m-%Y")

ok, but now i cant transform the data frame to a xts

[Felipe1]

it doesn't work with the format %d-%m-%Y or %Y%-%m-%d

[FJCC]

I have never used the xts() function, so treat my advice with care.
The documentation of the xts() function shows the following

xts(x = NULL,
    order.by = index(x),
    frequency = NULL,
    unique = TRUE,
    tzone = Sys.getenv("TZ"),
    ...)

The order.by argument is described as "a corresponding vector of unique times/dates".
One thing you have tried is

xts_price <- xts(price$crypto, price$price, price$date)

That assigns price$price to the order.by argument and since price$price is not a "vector of unique times/dates", you get the error that "order.by requires an appropriate time-based object".
Your second attempt is

    xts_price2 <- xts(select(price,-c("crypto","date")), 
                       order.by = as.POSIXct(strptime(price$date,"%Y-%m-%d")))

where you are trying to transform price$date. But price$date is already a Date object and does not need to be transformed. My guess - remember I have never used this function - is

xts_price2 <- xts(x = price$price, order.by = price$date)

[FJCC]

And looking at the image of your data, I would guess that you need to work with a single value of the Crypto column, but I really do not know what you are trying to do and I do not work with time series data.