Closed mdshw5 closed 8 years ago
@mdjones pivoted tables are only slightly more readable (maybe less) than the "melted" table:
>>> df.pivot('index', 'columns', 'values')
columns 100_q05 100_q25 100_q50 100_q75 100_q95 10_q05 10_q25 10_q50 \
index
UHR1_1_1 2 30 34 35 35 32 37 39
UHR1_1_2 2 25 33 35 35 30 37 39
columns 10_q75 10_q95 ... read_gc_91 read_gc_92 read_gc_93 \
index ...
UHR1_1_1 39 39 ... 40 23 20
UHR1_1_2 39 39 ... 52 44 29
columns read_gc_94 read_gc_95 read_gc_96 read_gc_97 read_gc_98 \
index
UHR1_1_1 14 8 6 2 2
UHR1_1_2 25 14 8 12 2
columns read_gc_99 reads
index
UHR1_1_1 1 103870017
UHR1_1_2 2 103870017
[2 rows x 4228 columns]
Making these tables more readable would require splitting the column names into key/value pairs and faceting sub-tables like:
read_gc counts
91 40
92 23
93 20
94 14
...
I don't actually remember the context for this request.
@mdjones prefers tab-separated values.