qsv to ods/xlsx makes strings out of numbers

[ondohotola]

I am extracting exchanges rates from a CSV provided by the EU Central Bank daily via
http://www.ecb.europa.eu/stats/eurofxref/eurofxref.zip

the extracted file of today eurofxref.csv contains

Date, USD, JPY, BGN, CZK, DKK, GBP, HUF, PLN, RON, SEK, CHF, ISK, NOK, TRY, AUD, BRL, CAD, CNY, HKD, IDR, ILS, INR, KRW, MXN, MYR, NZD, PHP, SGD, THB, ZAR, 
14 April 2025, 1.1377, 162.97, 1.9558, 25.108, 7.4675, 0.86383, 410.45, 4.2860, 4.9775, 11.0065, 0.9329, 144.90, 12.0205, 43.2931, 1.8034, 6.6488, 1.5786, 8.3196, 8.8221, 19125.14, 4.1910, 97.8348, 1617.29, 22.9720, 5.0229, 1.9430, 64.868, 1.4971, 38.176, 21.5332, 

qsv to ods eurofxref.ods eurofxref.csv shows all exchange rates as strings

same if I export to xlsx or sqlite, didn't try the others.

I expected those fields to be numerical.

• OS: Mac Squoia 13.4
• qsv Version 4.0.0

[jqnatividad]

If I'm not mistaken, ODS is schema-less and that may be why they are saved as strings.

One can run stats against the exported CSV and use the type info to cast it to the desired type, perhaps using luau.

Another workaround is to use the "smart" tojsonl command as it does this type-inferencing automatically, and then convert the jsonl back to csv with the jsonl command.

cc @kindly as he maintains the csvs_convert crate which powers the to command.

[kindly]

@ondohotola as the file contains spaces after the , it gets detected as a string, not a number. The conversion does this in order to give an accurate copy of the CSV, as in some cases the leading space might want to be kept.

You probably want to run qsv apply operations ltrim to strip the leading spaces first.

[ondohotola]

Indeed, thank you so much, I used

qsv input --trim-fields --trim-headers eurofxref.csv

to strip however :-0-O