csvlook: Change truncation character for floats

[csghone]

max-column-width does not work on columns interpreted as float.

eg:

echo "2.132456,156548.54646 2.132456,156548.54646" | xargs -n1 |csvlook --max-column-width 10
| 2.132456 | 156548.... |
| -------- | ---------- |
|   2.132… | 156,548... |

Expected behavior (or Workaround):

echo "2.132456,156548.54646 2.132456,156548.54646" | xargs -n1 |csvlook -I --max-column-width 100
| 2.132456 | 156548.54646 |
| -------- | ------------ |
| 2.132456 | 156548.54646 |

[jpmckinney]

Are you saying that the suggested behavior is to ignore --max-column-width for float columns? 156548.54646 is 12 characters long.

[csghone]

Sorry. My mistake.
Actually wanted to avoid the ellipsis character generated due to truncation - is there a way to specify a custom truncation character character? I think agate has one, but csvkit doesn't.

Do you prefer to close this ticket and create a new one? Or rename this one.

[jpmckinney]

I renamed the issue. We can implement this by exposing the max_precision parameter of agate's print_table as a command-line argument.

[jwocode]

the behaviour changed from csvkit==0.9.1 to csvkit==1.0.2
csvkit==0.9.1 was perfect for me getting exchange-rates from the European Bank. Also the missing column warning did not appear.

with the --no-inference option you can get close to the prior behavior.

before

$ csvsql  --query "select Date,USD,GBP,PLN,JPY from eurofxrefhist  where Date='2016-10-21'" eurofxrefhist.csv   |csvlook 
|-------------+--------+--------+--------+---------|
|  Date       | USD    | GBP    | PLN    | JPY     |
|-------------+--------+--------+--------+---------|
|  2016-10-21 | 1.0886 | 0.8932 | 4.3251 | 112.96  |
|-------------+--------+--------+--------+---------|

now

$ csvsql  --query "select Date,USD,GBP,PLN,JPY from eurofxrefhist  where Date='2016-10-21'" eurofxrefhist.csv   |csvlook 
/usr/lib/python3/dist-packages/agate/utils.py:275: UnnamedColumnWarning: Column 42 has no name. Using "qq".
|       Date |    USD |    GBP |    PLN |    JPY |
| ---------- | ------ | ------ | ------ | ------ |
| 2016-10-21 | 1,089… | 0,893… | 4,325… | 112,96 |