Max Harlow

Feature request: header detection

10

I know Python's `csv` library does this, would be great to have here and remove the need to specify `-H` when there's a header row. Additionally, there could perhaps be...

Feature request: delimiter detection

4

Would be nice not needing to specify `-d,` each time.

Output quoting doesn't work with newlines

3

Example CSV: ``` columnOne,columnTwo,columnThree testA,one one,aaa testB,"two two",bbb testC,"three, three",ccc ``` Running: ``` bash $ q -HOd, "select * from test.csv" ``` ``` columnOne,columnTwo,columnThree testA,one one,aaa testB,two two,bbb testC,"three, three",ccc...

Left and right join options for csvlink

2

We already have `--inner_join`, but would be useful to also have `--left_join` and `--right_join` too for when we want to also return items in one file not matched to the...

Option to specify fields by number

1

Some CSV files don't have headers, and it's somewhat awkward to add them just to use `csvlink`. So something like: ``` bash $ csvlink file1.csv file2.csv --field_numbers_1 3,4 --field_numbers_2 1,2...

Add other fuzzy matching algorithms

15

There's a list here: http://ntz-develop.blogspot.co.uk/2011/03/fuzzy-string-search.html

One option to specify fields to use to create the blocks, another (?) to set the method -- default to exact match, options for Metaphone etc Interesting bit on Soundex...

Investigate speed/memory improvements

3

eg: * generators and generator comprehensions * `cStringIO`, and other C-versions * streaming output? (probably required before #11) * write some performance tests

Speed up with memory mapping?

1

Could use something like Apache Arrow so files don't have to be kept in memory?

A la https://news.ycombinator.com/item?id=18613806