Is there a transformer that can convert a number into a number with a thousands separator, that is, "1000" to "1,000"? I know you can do it with Python but I can't find a transformer that does this.
Thanks
Page 1 / 1
Hi @jimo, I don't think a single transformer can do that but you can use a series of these transformers here. Assuming that an attribute called "_num" has a numeric value with integer representation:
StringPadder: Pad 0-2 spaces to the left side of the value of "_num" so that the number of characters will be a multiple of 3.
- Attributes: _num
- Desired String Length: @ceil(@StringLength(@Value(_num))/3.0)*3
- Side to Pad: Left
StringReplacer: Replace every three digits with the original three digits and a comma.
- Attributes: _num
- Text to Match: (...)
- Replacement Text: \\1,
- Use Regular Expressions: yes
AttributeTrimmer: Trim excess leading spaces and trailing comma from the resulting string.
- Attributes to Trim: _num
- Trim Type: Both
- Trim Characters: , <space and comma>
I see you specifically asked for a non-Python solution, but since it involves several transformers, here's one anyway ;-)
import fmeobjects
def FeatureProcessor(feature):
value = feature.getAttribute('num') or ''
try:
value_str = "{:,}".format(value)
except ValueError, TypeError:
value_str = value
feature.setAttribute('num', value_str)
Assumes input features with an attribute called "num" (change lines 4 and 9 as necessary)
Hi @jimo, I don't think a single transformer can do that but you can use a series of these transformers here. Assuming that an attribute called "_num" has a numeric value with integer representation:
StringPadder: Pad 0-2 spaces to the left side of the value of "_num" so that the number of characters will be a multiple of 3.
- Attributes: _num
- Desired String Length: @ceil(@StringLength(@Value(_num))/3.0)*3
- Side to Pad: Left
StringReplacer: Replace every three digits with the original three digits and a comma.
- Attributes: _num
- Text to Match: (...)
- Replacement Text: \1,
- Use Regular Expressions: yes
AttributeTrimmer: Trim excess leading spaces and trailing comma from the resulting string.
- Attributes to Trim: _num
- Trim Type: Both
- Trim Characters: , <space and comma>
@Trim(@ReplaceRegEx(@PadLeft(@Value(_num),@Evaluate(@ceil(@StringLength(@Value(_num))/3.0))*3),(...),"\1,")," ,") I think the stringformatter is the way to go.
.4f will be the format in your case
then a stringreplacer to replace the . to ,
Good luck
Those are all good answers, folks. I was really hoping that there was a single transformer that could do the job. I guess it's a good candidate for a "New Idea".
I try to limit the Python in my scripts because the next person who comes along might not know Python. However, the "{:,}".format(value) suggestion by david_r above seems to be the simplest in the Python realm.
Thanks
or ....
if you wanted a single transformer solution, you could use a bit of regex in a string replacer.
this will allow you to use one transformer against multiple attributes on the same feature, although admittedly many people may find the python script easier to maintain than the regex! neat solution though.
in 2015.0
StringReplacer settings
text to match:([0-9])(?=([0-9]{3})+$)
replacement text: \\1,
use reg expression: yes
if attr value does not match: no action
The 'positive lookahead assertion' regex is a modification of the version I found here , someone far cleverer than me worked it out. In my quick tests it worked on 1, 1234, 123456,123456789 adding the multiple commas in the right place if required.
or ....
if you wanted a single transformer solution, you could use a bit of regex in a string replacer.
this will allow you to use one transformer against multiple attributes on the same feature, although admittedly many people may find the python script easier to maintain than the regex! neat solution though.
in 2015.0
StringReplacer settings
text to match:([0-9])(?=([0-9]{3})+$)
replacement text: \\1,
use reg expression: yes
if attr value does not match: no action
The 'positive lookahead assertion' regex is a modification of the version I found here , someone far cleverer than me worked it out. In my quick tests it worked on 1, 1234, 123456,123456789 adding the multiple commas in the right place if required.
I see you specifically asked for a non-Python solution, but since it involves several transformers, here's one anyway ;-)
import fmeobjects
def FeatureProcessor(feature):
value = feature.getAttribute('num') or ''
try:
value_str = "{:,}".format(value)
except ValueError, TypeError:
value_str = value
feature.setAttribute('num', value_str)
Assumes input features with an attribute called "num" (change lines 4 and 9 as necessary)
https://knowledge.safe.com/questions/40697/format-number-from-1000000-to-1-000000.html
or ....
if you wanted a single transformer solution, you could use a bit of regex in a string replacer.
this will allow you to use one transformer against multiple attributes on the same feature, although admittedly many people may find the python script easier to maintain than the regex! neat solution though.
in 2015.0
StringReplacer settings
text to match:(/0-9])(?=(n0-9]{3})+$)
replacement text: \1,
use reg expression: yes
if attr value does not match: no action
The 'positive lookahead assertion' regex is a modification of the version I found here , someone far cleverer than me worked it out. In my quick tests it worked on 1, 1234, 123456,123456789 adding the multiple commas in the right place if required.
-12345566654.33345
-12345566654.33,345Just a heads up, this regex won't work for fractional numbers, only integers. Sample input:
-12345566654.33345
-12345566654.33,345My data happens to be integers in this instance, but I imagine there is away to incorporate(ignore) the fractions into the regex?
I will just stick an attributerounder in front in my case though!
Thanks @david_r, yes, good spot.
My data happens to be integers in this instance, but I imagine there is away to incorporate(ignore) the fractions into the regex?
I will just stick an attributerounder in front in my case though!
https://knowledge.safe.com/questions/40697/format-number-from-1000000-to-1-000000.html
I see you specifically asked for a non-Python solution, but since it involves several transformers, here's one anyway ;-)
import fmeobjects
def FeatureProcessor(feature):
value = feature.getAttribute('num') or ''
try:
value_str = "{:,}".format(value)
except ValueError, TypeError:
value_str = value
feature.setAttribute('num', value_str)
Assumes input features with an attribute called "num" (change lines 4 and 9 as necessary)
In my case, using python 3.6 it wont work due to ValueError, TypeError, If you remove these it works again.
import fmeobjects
def FeatureProcessor(feature):
value = feature.getAttribute('originalValue') or ''
try:
value_str = '{:,}'.format(value)
except:
value_str = value
feature.setAttribute('originalValue', value_str)