- Revision: the standalone
countcommand is replaced withlen, so make sure to replace(count)andcol "count"withlenandcol "len"respectively.- the unary
count <col>command is unaffected.
- the unary
cargo install dfsql
dfsql --input your.csv --output a-new.csv
# ...or
dfsql -i your.csv -o a-new.csvexit/quit: exit the REPL loop.exitundo: undo the previous successful operation.undo
reset: reset all the changes and go back to the original data frame.reset
schema: show column names and types of the data frame.schema
save: save the current data frame to a file.save a-new.csv
selectselect <expr>*
select last_name first_name- Select columns "last_name" and "first_name" and collect them into a data frame.
- Group by
group (<col> | <var>)* agg <expr>*
group first_name agg (count)
- Group the data frame by column "first_name" and then aggregate each group with the count of the members.
filterfilter <expr>
filter first_name = "John"
limitlimit <int>
limit 5
reversereverse
sortsort ((asc | desc | ()) <col>)*
sort icpsr_id
useuse <var>
use other
- Switch to the data frame called
other.
- Switch to the data frame called
- join
(left | right | inner | full) join <var> on <col> <col>?
left join other on id ID
- left join the data frame called
otheron my columnidand its columnID
- left join the data frame called
col: reference to a column.col : (<str> | <var>) -> <expr>
select col first_nameexclude: remove columns from the data frame.exclude : <expr>* -> <expr>
select exclude last_name first_name- literal: literal values like
42,"John",1.0, andnull. - binary operations
select a * b
- Calculate the product of columns "a" and "b" and collect the result.
- unary operations
select -a
select sum a- Sum all values in column "a" and collect the scalar result.
alias: assign a name to a column.alias : (<col> | <var>) <expr> -> <expr>
select alias product a * b
- Assign the name "product" to the product and collect the new column.
- conditional
<conditional> : if <expr> then <expr> (if <expr> then <expr>)* otherwise <expr> -> <expr>
select if class = 0 then "A" if class = 1 then "B" else null
cast: cast a column to either typestr,int, orfloat.cast : <type> <expr> -> <expr>
select cast str id- Cast the column "id" to type
strand collect the result.
- Cast the column "id" to type