orderby: Order By
In tera-insights/gtBase: R Interface for the Grokit System

Description Usage Arguments Details Value AUTO Author(s) Examples

The data is sorted according to a given ordering.

1
2
3

OrderBy(data, ..., inputs = AUTO, outputs = AUTO)

OrderByMake(..., limit = 0, rank = NULL)

`data`	an object of class `"data"`.
`inputs`	which attributes of the `data` to include in the result in addition to those used in the ordering.
`outputs`	possible re-namings of the .
`limit`	the number of rows to include in the result. A value of `0` specifies that all rows are kept. Setting `limit = k` effectively functions as a top-k selection.
`rank`	if non-`NULL`, an additional column is included that functions as the row number. This is typically not needed as the row number is otherwise provided. The name of this column is the value provided as `rank`.
`...`	For `OrderBy`: additional arguments passed to OrderByMake. For `OrderByMake`: a list of natural orderings to use, with precedence going to those seen earlier in the list. In the case that names are provided, these names are used as column names in the result. See ‘details’ for more information.

The ordering schema is provided as a list of arguments, each of which is in the form fun(expr) where fun is either asc or dsc and expr is an expression.

Precedence is based on the order in which the arguments are specified. For example, in the case of asc(att1), dsc(att2), the ordering is primarily ascending with respect to att1 with ties being broken by att2 in a descending manner. Ties are broken arbitrarily, including the case in which only some tuples that are tied are included in the result based on limit.

Each ordering expression is included in the result. If a name is provided in the argument list for a given expression, the column is given that corresponding name. Otherwise, if the expression is a single attribute then the column is given that attribute name. If not, then the column for that expression is given a constructed name that is hidden from the user and guaranteed to not conflict with other column names.

An object of class "data", with the attribute names as discussed above. Upon conversion to a data frame, there will be max(l, n), where l is limit and n is the number of rows in data.

In the case that inputs = AUTO, each attribute of the data that was not used expressly as an ordering attribute is included in the result. For example, if data contains attributes att1, att2, att3 and the ordering is asc(att1), dsc(att2 + att3), then the result will contain 4 columns with names att1, gen, att2, att3, where gen is a placeholder for a generated name and whose values are att2 + att3.

Jon Claus, <jonterainsights@gmail.com>, Tera Insights LLC

## TPCH Query 1
data <- Read(lineitem10g)
filter <- data[l_shipdate <= .(as.Date("1998-12-01")) - 90]
agg <- GroupBy(
  filter,
  groupAtts = c(rf = l_returnflag, ls = l_linestatus),
  sum_disc_price = Sum(l_extendedprice * (1 - l_discount)),
  sum_charge = Sum(l_extendedprice * (1 - l_discount) * (1 + l_tax)),
  avg_qty = Average(l_quantity),
  count_order = Count(1),
  sum_qty = Sum(l_quantity),
  avg_price = Average(l_extendedprice),
  sum_base_price = Sum(l_extendedprice),
  avg_disc = Average(l_discount)
)
agg <- OrderBy(
  agg,
  asc(rf),
  dsc(ls),
  rank = rank
)
result <- as.data.frame(agg)