Let’s add some tracing so we can see order-of-operations:
import sys
def findIdx(ary, tgt):
retval = ary.index(tgt)
print('First instance of %r in %r is at position %r' % (tgt, ary, retval), file=sys.stderr)
return retval
data1 = [1.48, -4.96]
i = 0
mn = data1[1]
k = findIdx(data1, mn)
data1[i], data1[k] = data1[k], data1[i]
print("Prior lookup: %r" % data1)
data2 = [1.48, -4.96]
data2[i], data2[findIdx(data2, mn)] = data2[findIdx(data2, mn)], data2[i]
print("Inline lookup: %r" % data2)
The logs are then elucidating:
First instance of -4.96 in [1.48, -4.96] is at position 1
Prior lookup: [-4.96, 1.48]
First instance of -4.96 in [1.48, -4.96] is at position 1
First instance of -4.96 in [-4.96, -4.96] is at position 0
Inline lookup: [1.48, -4.96]
As you can see, the last call to lookup the index is doing so against an array that has already had one of its two elements replaced.
1
solved Why does looking up an index *before* a swap rather than inline change the result?