`dcast()` test `2365.2` fail on R 3.5.0

[manmita]

As pointed by @aitap in #7634 which fixes #7614

Test 2365.2 ran without errors but failed check that x equals y:
> x = dcast(df_dcast, a ~ b, value.var = "v") 
   a  1  2  [Key=a Types=int,int,int Classes=fac,int,int]
1: x  3 NA                                               
2: y NA  4                                               
> y = dcast(dt_dcast, a ~ b, value.var = "v") 
   a  1  2  [Key=a Types=cha,int,int Classes=cha,int,int]
1: x  3 NA                                               
2: y NA  4                                               
Datasets have different column modes. First 3: a(numeric!=character)

The output of dcast is different in type for data.table and data.frame as data.frame converts char to factors for R version < 4.0.0
And dcast returns same output type as it was there in input hence we get column type mismatch.

[manmita]

The soln -> we can use string as factor as false for the test so that the type is same

df <- data.frame(
  name = c("Alice", "Bob"),
  age = c(25, 30),
  stringsAsFactors = FALSE
)

The test

df_dcast = data.frame(a = c("x", "y"), b = 1:2, v = 3:4)
dt_dcast = data.table(a = c("x", "y"), b = 1:2, v = 3:4)
test(2365.2, dcast(df_dcast, a ~ b, value.var = "v"), dcast(dt_dcast, a ~ b, value.var = "v"))

I think the dcast behavior is correct for data.frame as it converts the char column to factors then the output should be same preserving the type. so comparing equality between them isnt correct.

hence I suggest to make the input in same format and test for equality rather than code change for this fix

@aitap please let me know if I am thinking correctly

[manmita]

and maybe we can write a message for the users regarding the same