Jq is able to correctly extract “file_id” from “index_files”.Ĭommand: less example.json | jq -r '. However, it is unable to deal with complex nested records, such as "index_files": [ Ĭommand: less example.json | jsonv md5sum,data_type,file_name,analysis.analysis_id | less Jsonv deals with simple entries quite well.įor example, "md5sum": "68dc4de8dcc516b08aea9ce16017c505", This webpage would crash if I try to upload it. However, when the file size is too big, say, when I get the metadata record for over 9,000 samples, the file size goes as big as 90Mb. Just select the columns of interest and done! For example, for the several sub-records (is it the right term?) in “downstream_analyses”, it will give easy-to-understand column names like:ĭownstream_analyses/0/output_files/0/file_nameĭownstream_analyses/1/output_files/0/file_name You just need to upload the json file, and it will output every field of the entries. If the file size is too big the online tool to handle, use either If the json file is quite small in size, use the online tool provided in So, what is the best way to convert json to cvs format? At first I thought it should be quite simple, but it still took me several hours.
I have to convert it to csv, or tab-delimited format before I can process it. The metadata is only available in json format ( example of two entries ). For example, I work with TCGA data from GDC data portal for a while.
Sometimes the data you gather is in json format. It might be a minor issue, but has troubled me for a while.
0 kommentar(er)
